Randomise reconnect timing

[Steve-Mcl]

closes #314

Description

To alleviate stresses on broker, a variation in the reconnect timing is introduced

The exact value may wish to be tweaked but I went with upto 2 seconds sooner and no more than 10s later than the original 15s fixed time

Related Issue(s)

#314

Checklist

• I have read the contribution guidelines
• Suitable unit/system level tests have been added and they pass
• Documentation has been updated
  • Upgrade instructions
  • Configuration details
  • Concepts
• Changes flowforge.yml?
  • Issue/PR raised on FlowFuse/helm to update ConfigMap Template
  • Issue/PR raised on FlowFuse/CloudProject to update values for Staging/Production

Labels

• Includes a DB migration? -> add the area:migration label

[knolleary]

I think we should also include a similar randomised delay on startup. Perhaps not as extreme as 15secs, but between 2-10 secs for eg.

[Steve-Mcl]

I think we should also include a similar randomised delay on startup. Perhaps not as extreme as 15secs, but between 2-10 secs for eg.

As the mqtt.connect call is synchronous and following events rely on their being a client object created, the straightforward solution to this is to add a random await before the call to agent.start

There is only 1 reference to the agent.start function (and it is already an async fn), so a simple awaited delay:

await new Promise(r => setTimeout(r, randomInt(20, 5000)));

before line 138:

this.mqttClient.start()

would be feasible.

[Steve-Mcl]

a simple awaited delay:

await new Promise(r => setTimeout(r, randomInt(20, 5000)));

before line 138:

this.mqttClient.start()

would be feasible.

Added in commit add randomised start call delay

As this affects the start up time, I have kept is small enough to be less bothersome (20ms ~ 5s)

[knolleary]

@Steve-Mcl looks good - but a failing lint check in the tests

[Steve-Mcl]

@Steve-Mcl looks good - but a failing lint check in the tests

🤦‍♂️ Thats what I get for doing it in GH

fixed in 691eae7 🤞

[Steve-Mcl]

and of course, tests now timeout

Timeout of 2000ms exceeded

fix incoming.