Document Alert lifecycle

[r0h4n]

No description provided.

[mbukatov]

We need to include description of:

• polling loops in different components (node-agent, gluster-integration, monitoring, notifier) along with option to configure them
• UI and dashboard refresh cycle
• lifecycle of each type of events (gluster native events, other events) with path the information travels through tendrl components

[julienlim]

In addition, we may should consider including the following:

• Consider prioritizing certain events/alerts for scenarios that can involve unrecoverable data or data loss situations.
• It would be good to explain the data flow or lifecycle as an alert traverses through these various services (so one knows what impact tweaking the various intervals may offer.
• What is implications for each of the different components, i.e. node-agent, gluster-integration, monitoring-integration, notifier and what they cover if you tweak the polling/synch interval
• Tendrl should have it default to what our suggested best practice
• Ideally provide guidance for demo / sandbox environments, minimum, and recommended requirements and set expectations accordingly on what one can roughly expect (e.g. best case to worse case).

@r0h4n @Tendrl/tendrl-qe @julienlim

[julienlim]

@r0h4n @cloudbehl See also https://github.com/Tendrl/documentation/wiki/Best-Practices-for-Response-Times-and-Latency in helping set appropriate expectations.

[julienlim]

@brainfunked

Is there an updated component view of Tendrl that reflects what is in the current releases (e.g. 1.5.4)? It's not posted anywhere that I can find in this repo.

With regards to the above topic on the alert lifecycle and how it traverses through the Tendrl system, can you put something together? It would be very useful if we have it documented for Tendrl users. Thank you!