[thaumasiotes]

> Just wanted to say sorry for the outage. Issue was triggered by an aggressive uptime monitor which was freaking out at some slower-than-usual response times mid-HN-hug.

What is the theory behind this? If response times are too slow, how is terminating service supposed to help? Who is it supposed to help?

[_0vzt]

The timeout for the requests issued by the uptime monitor was set a little too low – the uptime monitor mistakenly believed the service was not responding (it was, but slowly)