Open luislhl opened 2 years ago
The issue HathorNetwork/hathor-wallet-headless#173 describes some weird behaviors when a wallet-headless app is trying to communicate with a heavily stressed fullnode, under test conditions. One of those is having very long response times for keeping the fullnode's websocket on, or even triggering connection timeouts like the one described here.
Maybe there is something in common between those two issues.
Could this be related to https://github.com/HathorNetwork/internal-issues/issues/370 ?
This is some random behavior we have observed while having a few wallet-headless connected to a fullnode.
The wallets are seldomly logging
Ping timeout. Connection is down... { uptime: 4.99, connectionTimeout: 5000 }
.We should investigate the reason for this, that probably lies in hathor-core, because the wallets and the fullnode in this case are running in the same network (with a ICMP ping latency of about 0.5ms).
One initial suspicion is that it may be related to CPU usage on the fullnode. We have observed a peak in CPU at the same time the wallets logged the timeout (all of them connected to the same fullnode logged this).