Merge ~cjwatson/launchpad-bastion-scripts:ignore-haproxy-queue-depth-checks into launchpad-bastion-scripts:main

[Colin Watson]

Ignore haproxy queue depth checks for production deployments

This doesn't mean that we should ignore them in general, just that they shouldn't block deployments.

[Ines Almeida (ines-almeida)]

How will we make sure we won't ignore them? Should we add in the documentation a note to check these specifically and raise them in the ~is chat or similar?

review: Approve

[Colin Watson (cjwatson)]

I think we'll need to update the nagiprom definitions in lp:canonical-is-prometheus. I filed https://warthogs.atlassian.net/browse/LP-1296 for this.

diff --git a/upgrade-production b/upgrade-production
index 9c6a5a1..da01057 100755
--- a/upgrade-production
+++ b/upgrade-production
@@ -1,5 +1,10 @@
 #! /bin/sh
 set -e
 
+# Ignore haproxy queue depth checks for deployment purposes: they may need
+# to be handled by IS, but are unlikely to relate very closely to code
+# deployments.
+skip_checks=check_haproxy_queue_depth
+
 mkdir -p "$HOME/logs"
-ssh [e-mail] autodeploy-trigger prod-launchpad 2>&1 | tee -a "$HOME/logs/$(date +%F).log"
+ssh [e-mail] autodeploy-trigger prod-launchpad EXTRA_SKIP_CHECKS="$skip_checks" 2>&1 | tee -a "$HOME/logs/$(date +%F).log"