feat: Add authenticated database readiness check based on driver session metadata - #2527
Open
erichare wants to merge 15 commits into
Open
feat: Add authenticated database readiness check based on driver session metadata#2527erichare wants to merge 15 commits into
erichare wants to merge 15 commits into
Conversation
Contributor
Unit Test Coverage Report
|
Contributor
Integration Test Coverage Report (dse69-it)
|
Contributor
Integration Test Coverage Report (hcd-it)
|
amorton
requested changes
Aug 4, 2026
amorton
left a comment
amorton
left a comment
Contributor
There was a problem hiding this comment.
-1 this only works in non-astra, uses blocking IO, and reads from a non replicated table so why the C* node will need to be marked as UP to respond it will not verify that its able to communicate with enough nodes to achieve quourm
need to rethink what we are trying to do here
Contributor
|
this needs a rethink about what we are trying to do |
Contributor
Author
understood! i'll address the specific comments you made just for the practice, but yeah understood about rethinking the purpose |
Contributor
Author
|
Okay, @amorton , so here's what i did... First of all, renamed the endpoint to The endpoint executes Switched everything to async, response would look like I'll push shortly |
erichare
changed the title
erichare
changed the title
Add check when creating a new CqlSession that it has Metadata so we know what keyspaces/tables in the DB. See comments in CqlSessionFactory
erichare
force-pushed
the
fix/2526-cassandra-readiness
branch
from
August 12, 2026 20:58
f9b504e to
aa3031a
Compare
Contributor
📈 Unit Test Coverage Delta vs Main Branch
|
Contributor
📈 Integration Test Coverage Delta vs Main Branch (dse69-it)
|
Contributor
📈 Integration Test Coverage Delta vs Main Branch (hcd-it)
|
Replace the datastax_sla.check canary query with a check of the driver session metadata: the pod is ready when the session obtained through the normal session cache reports at least one UP node. No query is issued and no canary table needs to be provisioned. Add stargate.jsonapi.operations.database-config.require-session-node-metadata (default false): when enabled, CqlSessionFactory rejects a newly built session whose metadata contains no nodes, closing it and failing creation with the standard FAILED_TO_CONNECT_TO_DATABASE error so an unusable session is never cached.
erichare
force-pushed
the
fix/2526-cassandra-readiness
branch
from
August 12, 2026 23:43
5be2ec1 to
e84e3c7
Compare
erichare
changed the title
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What this PR does:
Adds
GET /v1/health/ready, an authenticated readiness endpoint for both Astra and Cassandra.CQLSessionCacheusing the request's tenant, token, and User-Agent, and reportsUPwhen the session's driver metadata has anUPnode. No query is issued and no canary table is needed — a pod that cannot connect fails session creation and reportsDOWN.stargate.jsonapi.operations.sla-user-agent) so probe sessions get the shorter SLA cache TTL. Fails closed if it is unset.{"status":"UP"}, 401 invalid auth, 403 wrong User-Agent, 503{"status":"DOWN"}. Fully async with a 5s timeout. Readiness polling is excluded from tenant request metrics.This also means a probe token that authenticates but cannot read the schema makes the endpoint report
DOWNrather than an auth error.Which issue(s) this PR fixes:
Fixes #2526
Validation:
SessionEvictionIntegrationTestextended to assert readinessUP,DOWN,UPacross a container outage.Checklist