In order to gracefully handle pod termination in mw-on-k8s, envoy needs:
- A draining script
- A way to set --drain-time-s and --drain-strategy envoy CLI options
In order to gracefully handle pod termination in mw-on-k8s, envoy needs:
Status | Subtype | Assigned | Task | ||
---|---|---|---|---|---|
Resolved | Clement_Goubert | T331609 Gracefully handle pod termination in mw-on-k8s | |||
Resolved | Clement_Goubert | T338014 Allow setting draining strategy and time in envoy image |
Change 926455 had a related patch set uploaded (by Clément Goubert; author: Clément Goubert):
[operations/docker-images/production-images@master] envoy: Add draining script and configuration
Change 926455 merged by Clément Goubert:
[operations/docker-images/production-images@master] envoy: Add draining script and configuration
Mentioned in SAL (#wikimedia-operations) [2023-06-06T09:11:07Z] <claime> Building production images - T338014
Setting the following environment variables now controls envoy's draining behaviour:
DRAIN_TIME_S
DRAIN_STRATEGY
In case a non-default admin port is used, the ADMIN_PORT environment variable has to be set for the drain-envoy.sh script to work.
Change 927997 had a related patch set uploaded (by Clément Goubert; author: Clément Goubert):
[operations/docker-images/production-images@master] envoy: Add connection tracking to drain-envoy.sh
Change 927997 merged by Clément Goubert:
[operations/docker-images/production-images@master] envoy: Add connection tracking to drain-envoy.sh