Skip to content

[Bug]: [Spanner Change Streams] Increased polling frequency for idle change streams causes increased Cloud Spanner load #38707

@scwhittle

Description

@scwhittle

What happened?

37718 introduced a low-latency polling mode for spanner change streams including a shorter time range for polls as well as treating spanner heartbeat responses as idleness which should stop polling. Both of these were supposed to disabled by default but the logic for enabling heartbeats was inverted. This means that by default change streams stop polling when a heartbeat response is received (after configured time of 2 seconds). Or if low latency mode is enabled, the end timestamp of the query is 1 second after now but heartbeats do not trigger stopping the poll.

This mostly affects pipelines with relatively idle change streams which would otherwise hang on polling the subscription for a longer period of time even as heartbeats were sent on the idle change stream rpc.

A fix is being made in #38695, current work around is to use 2.72 or earlier release.

Issue Priority

Priority: 2 (default / most bugs should be filed as P2)

Issue Components

  • Component: Python SDK
  • Component: Java SDK
  • Component: Go SDK
  • Component: Typescript SDK
  • Component: IO connector
  • Component: Beam YAML
  • Component: Beam examples
  • Component: Beam playground
  • Component: Beam katas
  • Component: Website
  • Component: Infrastructure
  • Component: Spark Runner
  • Component: Flink Runner
  • Component: Prism Runner
  • Component: Twister2 Runner
  • Component: Hazelcast Jet Runner
  • Component: Google Cloud Dataflow Runner

Metadata

Metadata

Assignees

Labels

Type

No fields configured for Bug.

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions