Can you please suggest some ways how can we trigger notification whenever any of my production site goes down(Not capturing the events) i.e. suppose we have 100 production sites which are live and snowplow enabled and because of some production release in any 1 site our snowplow js is not working and we are not able to track the events on that site. As we can’t monitor 100 sites at all times, so want to know if there is any way to achieve this.
One way which I have is to compare the records count of every site today vs yesterday, if the record count goes below the threshold then we can trigger some kind of notification to the user.
So, we are using clojure collector with redshift db and tracking only web events using javascript tracker.
a) Events aren’t sending at all because of some unknown issue/change?
b) Fewer events are being sent because of a front-end Javascript change?
c) The same number of events are being sent but not sinking to Redshift (failing enrichment/validation)?
Recently I had came across the first event i.e. Events aren’t sending at all because of some unknown issue/change. Our production team had made some changes and we are not able to track any event.
All the above concerns are valid and user can came across any of the above. But the major one is first in which we are not able to track any events and want to generate notification for that.
You could set up cloud watch alerts for Kinesis raw (first stream after collector) having no data for X timeframe - then have it email you or pagerduty/slack whatever to alert you.
Thanks for your reply. But we are using clojure collector with ETL runner and we have more than 25 sites. So, can you please suggest a way to achieve the same in the above mention scenario.
Please let me know if you need more clarification.
