feature request: run doctor by default

[bam80]

The idea is:

• mtg runs doctor by default, before the usual start (can be opted out with a cli and/or config option)
• if the check fails, the start is aborted

@dolonet

[dolonet]

Thanks for the suggestion, but I'm going to close this as won't-implement. Reasoning:

mtg doctor is intentionally a pre-deploy diagnostic, not a startup gate. Looking at what it actually does (internal/cli/doctor.go):

• NTP query against 0.pool.ntp.org (external UDP, not a dependency mtg otherwise has)
• TCP dial to all 6 Telegram DCs (~10s worst case after doctor: run DC connectivity probes in parallel #485, longer if a proxy chain is configured)
• Fronting-domain reachability (10s)
• DNS lookup of the secret host + public-IP autodetect over HTTP
• PR doctor: deepen DC verification with MTProto handshake probe #496 adds a full MTProto handshake on top

Running this before every start would:

1. Add seconds of startup latency, plus several external dependencies (NTP pool, IP-echo service, DNS) that mtg run itself does not require.
2. Make mtg unstartable whenever any of those externals are flaky — including UDP-blocked Docker networks where NTP simply can't reach the pool.
3. Turn transient network blips into restart loops under restart: unless-stopped or systemd Restart=always — Docker/systemd backoff softens this but doesn't eliminate it.

The supported pattern is the one in the README: run mtg doctor config.toml once during deploy/config change, then start mtg run normally. CI or your deploy script is the right place to fail-fast on misconfiguration; the proxy hot path shouldn't depend on NTP or an IP-echo service to come up.

If there's a narrower variant you'd want — e.g. a startup-time config-only validation that does no network I/O — that's worth a separate issue with a concrete check list. The current mtg run already rejects malformed config at load; anything beyond that needs scoping.