Running out of DB connections on fxa-dev boxes?

[shane-tomlinson]

When running functional tests against fxa-ci we often see "Internal Server Error" in the test logs, which usually corresponds to an app.Translate related error in the auth-server logs.

Today when I ran the ansible script on fxa-ci, this error was reported (w/ host, user and password removed):

TASK: [sync | create syncserver database] *************************************
failed: [localhost] => {"changed": false, "cmd": ["mysql", "--host=XXX", "--user=XXX", "--password=XXX", "-e", "CREATE DATABASE IF NOT EXISTS sync"], "delta": "0:00:00.007519", "end": "2016-12-14 21:28:45.013063", "item": "", "rc": 1, "start": "2016-12-14 21:28:45.005544"}
stderr: ERROR 1040 (08004): Too many connections

FATAL: all hosts have already failed -- aborting

I tried several more times and kept seeing the same error. When I restarted the local fxa-auth-db-server, everything completed as expected.

Is the db-server holding open connections to the DB which would cause problems downstream?

[vladikoff]

stable and latest got more RDS connections! thanks to @jrgm