Files
neon/compute_tools/src/http/openapi_spec.yaml
Alexey Kondratov dd4fd89dc6 [compute_ctl] Do not initialize last_active on start (#4137)
Our scale-to-zero logic was optimized for short auto-suspend intervals,
e.g. minutes or hours. In this case, if compute was restarted by k8s due
to some reason (OOM, k8s node went down, pod relocation, etc.),
`last_active` got bumped, we start counting auto-suspend timeout again.
It's not a big deal, i.e. we suspend completely idle compute not after 5
minutes, but after 10 minutes or so.

Yet, some clients may want days or even weeks. And chance that compute
could be restarted during this interval is pretty high, but in this case
we could be not able to suspend some computes for weeks.

After this commit, we won't initialize `last_active` on start, so
`/status` could return an unset attribute. This means that there was no
user activity since start. Control-plane should deal with it by taking
`max()` out of all available activity timestamps: `started_at`,
`last_active`, etc.

compute_ctl part of neondatabase/cloud#4853
2023-05-05 11:45:37 +02:00

249 lines
6.4 KiB
YAML

openapi: "3.0.2"
info:
title: Compute node control API
version: "1.0"
servers:
- url: "http://localhost:3080"
paths:
/status:
get:
tags:
- Info
summary: Get compute node internal status.
description: ""
operationId: getComputeStatus
responses:
200:
description: ComputeState
content:
application/json:
schema:
$ref: "#/components/schemas/ComputeState"
/metrics.json:
get:
tags:
- Info
summary: Get compute node startup metrics in JSON format.
description: ""
operationId: getComputeMetricsJSON
responses:
200:
description: ComputeMetrics
content:
application/json:
schema:
$ref: "#/components/schemas/ComputeMetrics"
/insights:
get:
tags:
- Info
summary: Get current compute insights in JSON format.
description: |
Note, that this doesn't include any historical data.
operationId: getComputeInsights
responses:
200:
description: Compute insights
content:
application/json:
schema:
$ref: "#/components/schemas/ComputeInsights"
/info:
get:
tags:
- Info
summary: Get info about the compute pod / VM.
description: ""
operationId: getInfo
responses:
200:
description: Info
content:
application/json:
schema:
$ref: "#/components/schemas/Info"
/check_writability:
post:
tags:
- Check
summary: Check that we can write new data on this compute.
description: ""
operationId: checkComputeWritability
responses:
200:
description: Check result
content:
text/plain:
schema:
type: string
description: Error text or 'true' if check passed.
example: "true"
/configure:
post:
tags:
- Configure
summary: Perform compute node configuration.
description: |
This is a blocking API endpoint, i.e. it blocks waiting until
compute is finished configuration and is in `Running` state.
Optional non-blocking mode could be added later.
operationId: configureCompute
requestBody:
description: Configuration request.
required: true
content:
application/json:
schema:
type: object
required:
- spec
properties:
spec:
# XXX: I don't want to explain current spec in the OpenAPI format,
# as it could be changed really soon. Consider doing it later.
type: object
responses:
200:
description: Compute configuration finished.
content:
application/json:
schema:
$ref: "#/components/schemas/ComputeState"
400:
description: Provided spec is invalid.
content:
application/json:
schema:
$ref: "#/components/schemas/GenericError"
412:
description: |
It's not possible to do live-configuration of the compute.
It's either in the wrong state, or compute doesn't use pull
mode of configuration.
content:
application/json:
schema:
$ref: "#/components/schemas/GenericError"
500:
description: |
Compute configuration request was processed, but error
occurred. Compute will likely shutdown soon.
content:
application/json:
schema:
$ref: "#/components/schemas/GenericError"
components:
securitySchemes:
JWT:
type: http
scheme: bearer
bearerFormat: JWT
schemas:
ComputeMetrics:
type: object
description: Compute startup metrics.
required:
- wait_for_spec_ms
- sync_safekeepers_ms
- basebackup_ms
- config_ms
- total_startup_ms
properties:
wait_for_spec_ms:
type: integer
sync_safekeepers_ms:
type: integer
basebackup_ms:
type: integer
config_ms:
type: integer
total_startup_ms:
type: integer
Info:
type: object
description: Information about VM/Pod.
required:
- num_cpus
properties:
num_cpus:
type: integer
ComputeState:
type: object
required:
- start_time
- status
properties:
start_time:
type: string
description: |
Time when compute was started. If initially compute was started in the `empty`
state and then provided with valid spec, `start_time` will be reset to the
moment, when spec was received.
example: "2022-10-12T07:20:50.52Z"
status:
$ref: '#/components/schemas/ComputeStatus'
last_active:
type: string
description: |
The last detected compute activity timestamp in UTC and RFC3339 format.
It could be empty if compute was never used by user since start.
example: "2022-10-12T07:20:50.52Z"
error:
type: string
description: Text of the error during compute startup or reconfiguration, if any.
example: ""
tenant:
type: string
description: Identifier of the current tenant served by compute node, if any.
example: c9269c359e9a199fad1ea0981246a78f
timeline:
type: string
description: Identifier of the current timeline served by compute node, if any.
example: ece7de74d4b8cbe5433a68ce4d1b97b4
ComputeInsights:
type: object
properties:
pg_stat_statements:
description: Contains raw output from pg_stat_statements in JSON format.
type: array
items:
type: object
ComputeStatus:
type: string
enum:
- empty
- init
- failed
- running
- configuration_pending
- configuration
example: running
#
# Errors
#
GenericError:
type: object
required:
- error
properties:
error:
type: string
security:
- JWT: []