I started a Thingy:91 in the morning. It reports some data every hour over CoAP/DTLS 1.2 CID. What irritates me is, that it shows quite a lot of send failures. Sure, that maybe cause by UDP, but using a lot more devices with other CoAP/DTLS 1.2 CID endpoint, I’m usually get only a very few failures.
if possible, you may check my device id `655bafe778b93cd14474d10a.
It reports 10 messages and 6 failures.
Before sending data to golioth, that device uses an other endpoint for 4 days, sending about 110 messages with only 1 failure. The SIM card is the same. Do you have some server-side logs for that device?
Let me add.
the 6 failures happens before 1:07:02 PM. After that, no additional failures occur.
1 Like
@achim.kraus thank you for reporting this! We’ll look into these send failures.
Just to mention:
If you don’t see something in the server side logs, don’t waste time.
I will then enable the client side logs and we will see, if it happens again.
1 Like
It seems, there are two points:
First, the DTLS session seems to get lost every couple of days. I think, “DTLS graceful restart” isn’t implemented/supported.
Seconds, my Thingy app frequently “calls the watchdog”, but that doesn’t happen on my other Thingies, which are connected to Californium. I guess, there is a very bad and nasty bug in my code, when the device tries to reconnect. The good point, it’s pretty reproducible every couple of days.
1 Like
First, the DTLS session seems to get lost every couple of days. I think, “DTLS graceful restart” isn’t implemented/supported.
@achim.kraus this is correct and something we are actively working on addressing, especially given our recent support for CIDs. I will keep you updated on our progress!
1 Like