today if the db is down (local or remote), the notification service can't send an alert.
a general alert, similar to "engine is down" scenario should be sent.
difference is we don't have the db, so need in the config a new parameter for list of email addresses to send notification on "db is down".
probably after X retries, every Y minutes.
Verified on is5,
As long as DB is down, endless "db down" notification are sent.
Is this intentionally, or doed it need to be fixed to be only a single notification?
It is intentional in the default configuration.
along with FAILED_QUERIES_NOTIFICATION_RECIPIENTS which defines where emails should be sent, another configuration parameter was defined:
FAILED_QUERIES_NOTIFICATION_THRESHOLD. from the conf file documentation:
# Send a notification email after first failure to fetch notifications,
# and then once every failedQueriesNotificationThreshold times.
# 0 or 1 means notify on each failure.
The default value is 30. together with the default value for INTERVAL_IN_SECONDS (120) we will get one message every hour in the default configuration(!).
Ilanit, I see we have the docs_scoped flag as '?' does it mean the appropriate person will check if this feature (specifically the parameters) should be documented?