Skip to main content

Resource types

A resource type is a category of billable thing — an instance, a volume, a load balancer — and each type exposes a set of attributes. Attributes are the vocabulary of the billing engine: every price, filter, modifier and adjustment rule target is written in terms of them.

The list below is what a fully-featured deployment exposes. Your own list comes from the cloud providers and service integrations you have configured — a deployment without shared file systems never reports share, and a registered billing resource exporter contributes its own types. The authoritative list for a given deployment is the Resource Type dropdown in a price plan rule.

Reading the tables​

Data type decides where an attribute can be used:

Data typeMeaningWhere it can be used
booleantrue / false, read as 1 / 0Filters and modifiers
numberNumeric measure, in the unit shown in bracketsPrices, filters, modifiers
stringText, e.g. a flavor name or a statusFilters and modifiers
mapKey/value pairs addressed one key at a time, e.g. metadataFilters and modifiers

Only numeric attributes can carry a price, because a price multiplies an amount by the attribute's value. existence is the exception and is offered on every type. Everything else narrows when a price applies, or adjusts it — to charge for a boolean condition, filter on the boolean and price existence.

Units matter, because the amount you enter is per unit of the attribute: size on a volume is GB, so 0.10 prices a volume at €0.10 per GB, while size on a bucket is bytes and the same 0.10 would price a gigabyte at over €100,000,000. Where the same name carries different units on different resource types, the tables below show the unit that resource type actually reports. Attributes whose name ends in _gb, _gib, _mb, _bytes or _hours state their unit in the name; bucket and image also expose size_gb alongside the raw byte count for exactly this reason.

Counter separates the two ways an attribute changes over time, which is what decides how the billing job writes the result to the bill:

  • A gauge (Counter: —) describes the resource as it is right now: vcpus, size, existence. Each job run adds its result to the bill line.
  • A counter (Counter: yes) is a total that has been accumulated for you, such as total_traffic_mb or gib_hours. Each job run overwrites the bill line with the latest total, because the number already includes everything billed so far this cycle.

A single rule cannot mix the two — see Gauges and counters.

Most attributes are the cloud's own value under the cloud's own name. Where an attribute name is a link, it is not — the link leads to how that one is resolved.

Sources of attribute values​

Nothing is fetched from a cloud while a bill is being rated. Every attribute is read from something OSIE already holds, and there are three ways it gets there.

The record kept for each cloud resource. Everything that describes a resource as it is now — its flavor, size, status, availability zone — comes from the copy OSIE keeps of that resource, refreshed by the periodic resource sync and by the cloud's own notifications. A resource the sync has not seen yet is not on the bill, and an attribute the cloud stopped reporting keeps its last known value until the next sync.

Usage collected per billing cycle. Counters are collected by a separate job that runs every 15 minutes and keeps one running total per resource per billing cycle. Instance traffic, object-storage traffic, backup storage held and managed-database usage all arrive this way. The figure is always cycle-to-date, which is why pricing a counter overwrites the bill line rather than adding to it, and why the resource types built out of it — instance_traffic, bucket_traffic, vprotect_backup_storage and the s9s_* types — carry counters and little else.

A billing resource exporter. A deployment can register an external service that reports its own resource types and attributes over the Billing Resource Exporter API. Those types appear in the Resource Type dropdown next to the built-in ones and are priced identically.

Attributes OSIE adds itself​

Three attributes are not reported by the resource at all. existence is on every type; service_id and region are on every type that belongs to a cloud provider.

existence​

existence is synthesised by OSIE, and it is always 1 while the resource is present in the customer's account. It is the attribute to price when you charge for a resource merely existing — "€0.10 per hour per instance". It is also the only price offered on every resource type, which makes it the way to charge for any condition a filter can express.

service_id and region​

OSIE attaches both to every resource it holds for a cloud provider, so you can filter on them to price the same resource type differently per cloud or per region within one price plan.

They are absent from the types that are not tied to one cloud resource — project and the types built from collected usage. A traffic or managed-database rule therefore cannot be narrowed per cloud or per region; restrict the whole price plan to a cloud provider instead, on its General tab.

Missing attributes​

A resource does not always report every attribute its type declares, and what happens then is not the same in a price as in a filter. The difference decides how a mistyped rule fails:

  • In a filter or a modifier, a missing attribute matches only the negations: not equals, and on a text attribute does not contain and does not start with. With any other operator the rule does not apply, so a filter on the wrong attribute, or on a metadata key that nothing sets, prices nothing. With not equals the rule does apply, which is what lets metadata.<key> not equals <value> exclude only the resources carrying that value.
  • In a price, a missing attribute counts as 0. The rule applies, and adds nothing.

Resource types​

Compute​

Virtual and physical machines are separate resource types, and a rule applies to one of them.

instance​

Source: nova / server

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
availability_zonestring—Availability zone the resource is placed in
instance_typestring—Flavor the instance runs on
ram_mbnumber (MB)—RAM of the flavor
ram_gbnumber (GB)—RAM of the flavor
statusstring—Power state, e.g. ACTIVE, SHUTOFF, ERROR
vcpusnumber (count)—vCPUs of the flavor
root_disk_gbnumber (GB)—Root disk of the flavor
hoststring—Hypervisor the instance is placed on
incoming_public_traffic_mbnumber (MB)yesInbound traffic from outside the cloud, this billing cycle
outgoing_public_traffic_mbnumber (MB)yesOutbound traffic to outside the cloud, this billing cycle
incoming_private_traffic_mbnumber (MB)yesInbound traffic within the cloud, this billing cycle
outgoing_private_traffic_mbnumber (MB)yesOutbound traffic within the cloud, this billing cycle
total_public_traffic_mbnumber (MB)yesInbound plus outbound public traffic, this billing cycle
total_private_traffic_mbnumber (MB)yesInbound plus outbound private traffic, this billing cycle
total_traffic_mbnumber (MB)yesAll traffic, public and private, this billing cycle
imagestring—Image the instance booted from
metadatamap—Instance and image metadata, addressed one key at a time
backup_enabledboolean—Whether a backup product protects the instance
backup_providerstring—Backup product protecting the instance
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider
metadata​

metadata is a map, so it holds keys that nothing registers in advance. The dropdown offers metadata once, you name the key next to it, and the rule is stored against that one entry — metadata.firewall, metadata.os. This is how a charge follows the workload deployed rather than the size of the machine running it. instance, bare_metal and volume carry it.

On an instance, two sets of key/value pairs are merged into it:

  • The metadata on the instance, set when it is created — either from the metadata options an administrator publishes or as free-form keys — and editable afterwards. It is read back from the cloud by the resource sync.
  • The properties of the image it booted from, read once per instance and kept alongside the record. For an instance booted from a volume, the image properties copied onto the boot volume are used instead, so one rule matches an instance whichever way it booted. The image's own id and name are additionally available as image_id and image_name.

On the same key, the image wins. Image properties are set by the operator on the image, while instance metadata is editable by the customer — what is billed must not be changeable by the party being billed.

Each key's data type is decided from its value at billing time: true or false gives a boolean, a numeric value gives a number, anything else is text. That decides which operators the value field offers.

backup_enabled and backup_provider​

Neither comes from the cloud. Both are resolved per project and cloud from the backup products configured on the platform: backup_enabled is true while any of them protects the instance, and backup_provider names the one that does. The backup_provider value list offers only the products this platform actually runs, so it cannot produce a rule that can never match.

They are the way to charge a backup licence for the instances that are actually being protected: filter on backup_enabled, then price existence. Every other dimension you would tier such a licence on — flavor, vCPUs, RAM, image, metadata — is on the same resource, so one rule can combine them. Only instance carries these two.

image​

image holds the id of the image the machine booted from, and its value field is a dropdown of the cloud provider's public images rather than a free-text box. A machine booted from a volume has no image of its own; the id recorded on its boot volume is used instead, so an image rule does not silently skip every boot-from-volume machine.

The traffic counters​

The *_traffic_mb attributes, here and on instance_traffic, are read from the metrics backend configured for that cloud and region — see Instance metrics. Where no metrics source is configured they are not reported at all, and a rule pricing them charges nothing.

Public and private are decided per network interface: traffic on an interface attached to an external network counts as public, everything else as private. The total_* attributes are sums of the directional ones, so pricing total_traffic_mb alongside incoming_public_traffic_mb charges some bytes twice.

bare_metal​

Source: ironic / server

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
availability_zonestring—Availability zone the resource is placed in
instance_typestring—Flavor the node was provisioned with
ram_mbnumber (MB)—RAM of the flavor
ram_gbnumber (GB)—RAM of the flavor
statusstring—Power state of the node
vcpusnumber (count)—vCPUs of the flavor
root_disk_gbnumber (GB)—Root disk of the flavor
hoststring—Physical node the server runs on
imagestring—Image the node was provisioned with
is_bareMetalboolean—Always true; separates bare metal from virtual instances
metadatamap—Node and image metadata, addressed one key at a time
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

Bare metal reports no traffic and no backup protection. metadata and image behave as they do on instance.

instance_traffic​

Source: Gnocchi or Prometheus / instance traffic

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
incoming_public_traffic_mbnumber (MB)yesInbound traffic from outside the cloud, this billing cycle
outgoing_public_traffic_mbnumber (MB)yesOutbound traffic to outside the cloud, this billing cycle
incoming_private_traffic_mbnumber (MB)yesInbound traffic within the cloud, this billing cycle
outgoing_private_traffic_mbnumber (MB)yesOutbound traffic within the cloud, this billing cycle
total_public_traffic_mbnumber (MB)yesInbound plus outbound public traffic, this billing cycle
total_private_traffic_mbnumber (MB)yesInbound plus outbound private traffic, this billing cycle
total_traffic_mbnumber (MB)yesAll traffic, public and private, this billing cycle

The same counters as instance carries, as a type of their own for pricing traffic separately from the machine. It has no other attributes, so a rule on it cannot be narrowed by flavor, cloud or region.

Block storage​

volume​

Source: cinder / volume

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
availability_zonestring—Availability zone the resource is placed in
sizenumber (GB)—Provisioned size
bootableboolean—Whether the volume can boot an instance
statusstring—Provider-reported state of the resource
display_namestring—Name shown for the resource
typestring—Volume type, which usually decides the storage tier
metadatamap—Volume and source-image metadata, addressed one key at a time
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider
metadata on volumes​

On a volume, metadata merges the volume's own metadata with the properties of the image the volume was created from, which OpenStack copies onto the volume when it creates it — the image's image_id and image_name among them. On the same key, the image wins. Both are read from the record the resource sync keeps, so existing volumes carry them from their next sync.

This is how volumes charged as part of a service are kept out of per-GB pricing: filter the volume rule on a key the service's image sets, with not equals. Volumes without the key still match, so volumes created empty or from other images keep being priced.

  • Customers cannot set a volume's metadata from the portal; it is set with the OpenStack API or CLI. OpenStack's default policy lets project members change both sets on their own volumes, so an exclusion customers must not be able to trigger needs a policy you have restricted.
  • The price a customer sees while creating a volume does not take metadata rules into account, so a rule on it can make the first bill differ from that estimate.

volume_snapshot​

Source: cinder / snapshot

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
sizenumber (GB)—Provisioned size
statusstring—Provider-reported state of the resource
display_namestring—Name shown for the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

volume_backup​

Source: cinder / backup

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
sizenumber (GB)—Provisioned size
statusstring—Provider-reported state of the resource
display_namestring—Name shown for the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

Object storage​

bucket​

Source: swift / container

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
sizenumber (bytes)—Stored data in the container
size_gbnumber (GB)—Stored data in the container
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

bucket_traffic​

Source: VHI object storage / usage report

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
uploaded_bytesnumber (bytes)yesData written to object storage, this billing cycle
downloaded_bytesnumber (bytes)yesData read from object storage, this billing cycle

Reported only where object storage is served by VHI and its usage reporting is configured — see Configure VHI S3.

Shared file systems​

share​

Source: manila / share

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
availability_zonestring—Availability zone the resource is placed in
sizenumber (GiB)—Provisioned size
share_typestring—Share type, which usually decides the storage tier
protocolstring—Access protocol, e.g. NFS or CIFS
statusstring—Provider-reported state of the resource
display_namestring—Name shown for the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

share_network​

Source: manila / share network

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
network_typestring—Network type backing the share network
statusnumber—Provider-reported state, reported as a number
display_namestring—Name shown for the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

share_security_service​

Source: manila / security service

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
typestring—Security service type, e.g. LDAP or Active Directory
display_namestring—Name shown for the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

share_snapshot​

Source: manila / snapshot

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
display_namestring—Name shown for the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

share_snapshot_group​

Source: manila / snapshot group

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
display_namestring—Name shown for the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

Networking​

network​

Source: neutron / network

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
has_router_externalboolean—Whether the network is an external, provider network
sharedboolean—Whether the network is shared with other projects
network_typestring—Segmentation type, e.g. VXLAN or VLAN
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

subnet​

Source: neutron / subnet

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
enabled_dhcpboolean—Whether DHCP is served on the subnet
ip_versionnumber—IP version of the subnet, 4 or 6
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

port​

Source: neutron / port

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
statestring—Administrative state of the port
network_idstring—Network the port belongs to
device_idstring—Resource the port is attached to
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

router​

Source: neutron / router

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
has_external_gatewayboolean—Whether the router has an external gateway set
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

floating_ip​

Source: neutron / floating IP

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
statusstring—Provider-reported state of the resource
floating_network_idstring—External network the address was allocated from
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

security_group​

Source: neutron / security group

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
rulesnumber (count)—Number of rules in the group
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

load_balancer​

Source: octavia / load balancer

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
flavor_idstring—Load balancer flavor, which decides its capacity
operating_statusstring—Operating status reported by the load balancer
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

firewall_group​

Source: neutron-fwaas / firewall group

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
portsnumber (count)—Number of router interfaces the group filters
display_namestring—Name shown for the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

firewall_policy​

Source: neutron-fwaas / firewall policy

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
display_namestring—Name shown for the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

firewall_rule​

Source: neutron-fwaas / firewall rule

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
actionstring—What the rule does with matching traffic: allow, deny or reject
display_namestring—Name shown for the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

VPN​

vpn_service​

Source: neutron-vpnaas / VPN service

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
flavor_idstring—VPN service flavor
subnet_idstring—Subnet the VPN service terminates on
display_namestring—Name shown for the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

vpn_endpoint_group​

Source: neutron-vpnaas / endpoint group

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
typestring—Endpoint group type, e.g. subnet or CIDR
display_namestring—Name shown for the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

ike_policy​

Source: neutron-vpnaas / IKE policy

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
encryption_algorithmstring—Encryption algorithm of the IKE policy
display_namestring—Name shown for the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

ipsec_policy​

Source: neutron-vpnaas / IPsec policy

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
encryption_algorithmstring—Encryption algorithm of the IPsec policy
display_namestring—Name shown for the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

ipsec_site_connection​

Source: neutron-vpnaas / site connection

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
peer_addressstring—Address of the remote peer
peer_idstring—Identifier of the remote peer
display_namestring—Name shown for the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

Images​

image​

Source: glance / image

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
image_typestring—Image type, e.g. snapshot or uploaded image
image_namestring—Name of the image
sizenumber (bytes)—Image size as stored
size_gbnumber (GB)—Image size as stored
os_distrostring—Operating system distribution
os_typestring—Operating system family
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

Orchestration​

stack​

Source: heat / stack

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
display_namestring—Name shown for the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

Kubernetes​

kubernetes_cluster​

Source: magnum / cluster

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
cluster_template_idstring—Template the cluster was created from
master_countnumber (count)—Control plane nodes in the cluster
node_countnumber (count)—Worker nodes in the cluster
master_lb_enabledboolean—Whether the control plane sits behind a load balancer
floating_ip_enabledboolean—Whether nodes get floating IPs
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

A cluster adds no charge while Magnum creates it, after it failed to create, or while Magnum deletes it or after a delete failed. A cluster that comes up is charged from its creation, build time included, as an instance is. A cluster that fails to create is never charged.

DNS​

dns​

Source: designate / zone

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
statusstring—Provider-reported state of the resource
dns_emailstring—Contact address on the zone
dns_ttlnumber (seconds)—Default record TTL of the zone
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

Key management​

barbican_secret​

Source: barbican / secret

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
display_namestring—Name shown for the resource
secret_typestring—Secret type, e.g. symmetric or certificate
statusstring—Provider-reported state of the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

barbican_container​

Source: barbican / container

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
display_namestring—Name shown for the resource
typestring—Container type, e.g. generic or certificate
statusstring—Provider-reported state of the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

Databases​

Two shapes of database service report here, and they are priced differently.

database_instance and database_backup are database resources running in the OpenStack cloud: one billable resource each, described as it is right now, priced like an instance or a volume.

The s9s_* types are the usage a managed database service reports for the cycle so far. There is no per-database resource to price — a single datastore contributes to several of these types at once, each carrying counters and the dimensions to filter them by. Price them in their own rules.

database_instance​

Source: trove / instance

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
flavor_idstring—Flavor the database instance runs on
volume_sizenumber (GB)—Provisioned storage of the instance
volume_typestring—Storage type backing the instance
datastorestring—Database engine, e.g. mysql or postgresql
datastore_versionstring—Engine version
statusstring—Provider-reported state of the resource
display_namestring—Name shown for the resource
availability_zonestring—Availability zone the resource is placed in
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

database_backup​

Source: trove / backup

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
sizenumber (GB)—Size of the backup
statusstring—Provider-reported state of the resource
display_namestring—Name shown for the resource
datastorestring—Database engine of the backed-up instance
datastore_versionstring—Engine version of the backed-up instance
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

s9s_instance_type_usage​

Source: Severalnines / instance usage

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
datastorestring—Database engine the usage belongs to
datastore_namestring—Name of the datastore instance
datastore_typestring—Type of the datastore
db_vendorstring—Database vendor
typestring—Instance type the usage is reported for
hoursnumber (hours)yesInstance hours accrued this billing cycle

s9s_volume_type_usage​

Source: Severalnines / volume usage

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
datastorestring—Database engine the usage belongs to
datastore_namestring—Name of the datastore instance
datastore_typestring—Type of the datastore
db_vendorstring—Database vendor
typestring—Volume type the usage is reported for
average_gibnumber (GiB)yesAverage provisioned storage this billing cycle
gib_hoursnumber (GiB-hours)yesStorage-hours accrued this billing cycle
iops_hoursnumber (IOPS-hours)yesIOPS-hours accrued this billing cycle

s9s_network_egress_usage​

Source: Severalnines / egress usage

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
datastorestring—Database engine the usage belongs to
datastore_namestring—Name of the datastore instance
datastore_typestring—Type of the datastore
db_vendorstring—Database vendor
gibnumber (GiB)yesEgress traffic this billing cycle

s9s_backup_usage​

Source: Severalnines / backup usage

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
datastorestring—Database engine the usage belongs to
datastore_namestring—Name of the datastore instance
datastore_typestring—Type of the datastore
db_vendorstring—Database vendor
removed_countnumber (count)yesBackups removed this billing cycle
removed_gibnumber (GiB)yesBackup storage freed this billing cycle
storage_gibnumber (GiB)yesBackup storage held this billing cycle
taken_countnumber (count)yesBackups taken this billing cycle

Backup​

Backup products report at different granularity. Trilio exposes the workloads and snapshots it holds as resources of their own, so they can be filtered and tiered individually. Storware vProtect reports a single storage figure per project and cloud, in both of its count modes, so the same protected data can be charged on what it occupies after deduplication or on its size before it.

Protection of an instance is charged from the instance itself, through backup_enabled.

trilio_workload​

Source: Trilio / workload

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
storage_usage_bytesnumber (bytes)yesBackup storage consumed by the workload
statusstring—Provider-reported state of the resource
encryptionboolean—Whether the workload is encrypted
instances_countnumber (count)—Instances protected by the workload
display_namestring—Name shown for the resource
backup_target_typesstring—Target types the workload backs up to
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

trilio_snapshot​

Source: Trilio / snapshot

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
sizenumber (bytes)—Size of the snapshot
statusstring—Provider-reported state of the resource
snapshot_typestring—Snapshot type, e.g. full or incremental
display_namestring—Name shown for the resource
service_idstring—Cloud provider the resource belongs to
regionstring—Region within the cloud provider

vprotect_backup_storage​

Source: Storware vProtect / storage

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
stored_gbnumber (GB)yesBackup data held in storage
frontend_gbnumber (GB)yesProtected source data, as seen before deduplication

Platform​

project​

Source: OSIE / project

AttributeData typeCounterDescription
existenceboolean—Always 1 while the resource exists in the account
display_namestring—Name of the project
billing_profile_idstring—Billing profile the project is billed to
project_idstring—Identifier of the project in OSIE

The project itself, billable independently of what it contains — a flat platform fee per project, or a fee that varies by billing profile.