telegram server automatically closes my TCP connection

[xiaolan123]

When I add many public groups or channels to my account, if there is a review message, the link will be automatically closed by the tg server every 15 minutes or so

tdlog:
{"level":"debug","ts":"2023-05-29 11:24:52","logger":"conn.mtproto.read","msg":"Read loop done","v":"v0.83.0","conn_id":0,"dc_id":5,"reason":"read: read: read intermediate: read length: EOF","reasonVerbose":"read:\n github.com/gotd/td/internal/mtproto.(*Conn).readLoop\n /root/go/pkg/mod/github.com/gotd/[email protected]/internal/mtproto/read.go:214\n - read:\n github.com/gotd/td/transport.(*connection).Recv\n /root/go/pkg/mod/github.com/gotd/[email protected]/transport/connection.go:70\n - read intermediate:\n github.com/gotd/td/internal/proto/codec.Intermediate.Read\n /root/go/pkg/mod/github.com/gotd/[email protected]/internal/proto/codec/intermediate.go:73\n - read length:\n github.com/gotd/td/internal/proto/codec.readLen\n /root/go/pkg/mod/github.com/gotd/[email protected]/internal/proto/codec/codec.go:35\n - EOF"}

tcpdump
91.108.56.125.443 > 103.122.98.115.57236: Flags [P.], cksum 0x042c (correct), seq 904965:906193, ack 35404, win 14912, options [nop,nop,TS val 149279870 ecr 107897976], length 1228
00:14:56.670365 IP (tos 0x0, ttl 64, id 25105, offset 0, flags [DF], proto TCP (6), length 52)
103.122.98.115.57236 > 91.108.56.125.443: Flags [.], cksum 0x5f56 (incorrect -> 0x5454), ack 906193, win 1424, options [nop,nop,TS val 107898013 ecr 149279870], length 0
00:14:56.671105 IP (tos 0x0, ttl 55, id 2473, offset 0, flags [DF], proto TCP (6), length 1200)
91.108.56.125.443 > 103.122.98.115.57236: Flags [P.], cksum 0xf885 (correct), seq 906193:907341, ack 35404, win 14912, options [nop,nop,TS val 149279871 ecr 107897976], length 1148
00:14:56.671142 IP (tos 0x0, ttl 64, id 25106, offset 0, flags [DF], proto TCP (6), length 52)
103.122.98.115.57236 > 91.108.56.125.443: Flags [.], cksum 0x5f56 (incorrect -> 0x4fd6), ack 907341, win 1424, options [nop,nop,TS val 107898014 ecr 149279871], length 0
00:14:56.706742 IP (tos 0x0, ttl 55, id 2474, offset 0, flags [DF], proto TCP (6), length 52)
91.108.56.125.443 > 103.122.98.115.57236: Flags [F.], cksum 0x1b01 (correct), seq 907341, ack 35404, win 14912, options [nop,nop,TS val 149279907 ecr 107898014], length 0
00:14:56.746132 IP (tos 0x0, ttl 64, id 25107, offset 0, flags [DF], proto TCP (6), length 52)
103.122.98.115.57236 > 91.108.56.125.443: Flags [.], cksum 0x5f56 (incorrect -> 0x4f66), ack 907342, win 1424, options [nop,nop,TS val 107898089 ecr 149279907], length 0
00:15:03.839441 IP (tos 0x0, ttl 64, id 25108, offset 0, flags [DF], proto TCP (6), length 56)

[xiaolan123]

  my code 

client.Run(ctx, func(ctx context.Context) error {
	for true {
		SendMessageRequest := &tg.MessagesSendMessageRequest{
			Peer: &tg.InputPeerChannel{
				ChannelID:  chatId,
				AccessHash: chatHash,
			},
			Message:  "hello",
			RandomID: randId,
		}
		client.API().MessagesSendMessage(ctx, SendMessageRequest)
		time.Sleep(30 * time.Second)
	}
	return nil
})

[xiaolan123]

client.Run(ctx, func(ctx context.Context) error {
for true {
time.Sleep(30 * time.Second)
}
return nil
})

I also found that even if I didn't do anything, I would still make this mistake ： read length: EOF

[jhemmmm]

Probably this?.
#1030

[ernado]

I'm closing this in favor of #1030.

I'm trying to find a way to reproduce this issue.
Tried something like that:

sudo ss -K dst 149.154.167.50

Just closing connection does not reproduce this issue, so probably another way is needed.