As seen in T198042, WDQS has a number of threads stuck on logging. We should use an async logger and drop messages when the logging throughput can't be reached.
Of course, we should also reduce the amount of logging, but that will be for another task.
| Status | Subtype | Assigned | Task | ||
|---|---|---|---|---|---|
| Resolved | None | T198042 WDQS timeout on the public eqiad cluster | |||
| Resolved | None | T198051 Enable async logging on Wikidata Query Service |
Change 441772 had a related patch set uploaded (by Gehel; owner: Gehel):
[operations/puppet@production] wdqs: enable async logging
Change 441772 merged by Gehel:
[operations/puppet@production] wdqs: enable async logging
Mentioned in SAL (#wikimedia-operations) [2018-06-25T08:11:23Z] <gehel> rolling restart of wdqs to enable async logging - T198051
Change 441781 had a related patch set uploaded (by Gehel; owner: Gehel):
[operations/puppet@production] wdqs: async appender only support a single appender-ref
Change 441781 merged by Gehel:
[operations/puppet@production] wdqs: async appender only support a single appender-ref
Change 441803 had a related patch set uploaded (by Gehel; owner: Gehel):
[operations/puppet@production] wdqs: correct async appender package name
Change 441803 merged by Gehel:
[operations/puppet@production] wdqs: correct async appender package name