solution for send failures and/or retries

[problame]

Ideas

• keep the last panic in a file in /var/lib/*; when sending the next message, try to send that panic first (but don't catch that panic)
• implement a local queue
  • sendmail mode is a special case that saves to /var/lib/ before entering common mode which...
  • ... for each file in /var/lib/, does one transmission attempt and deletes the file on successful transmission (opening a small window for duplicate transmission, that's ok)
  • main problem with a queue: what to do to avoid it overflowing
• fallback notification mechanism => healthchecks.io allows pinging on failures