If chasquid runs out of file descriptors, that can cause problems accepting connections, writing to the queue, and even reloading configs. Some of those errors are fatal, as they should not happen under normal conditions. Since a remote client can open a lot of connections, this can end up consuming all file descriptors, and resulting in a denial of service. To reduce the impact of that kind of event, this patch makes chasquid limit the number of concurrent SMTP connections to half the maximum number of file descriptors. Connections over that limit will get closed immediately. The server will log (rate-limited) when such event happens, and new metrics make these events visible through monitoring too. This approach complements the previous patch, which made the server retry Accept errors, to reduce the impact of connection overloads. This mainly affects deployments with low file descriptor limits; most modern deployments have very high limits by default, so they are unlikely to experience this problem in the first place. Users who have to be very defensive can deploy OS-level safeguards (such as firewall rules to limit number of connections, etc.) which are more effective and tunable. Thanks to John JEAN <john@wargan.com> for finding and reporting this problem. The detailed report can be found in https://johnjean.io/advisories/chasquid-unauthenticated-remote-dos/.
5.4 KiB
Monitoring
chasquid includes an HTTP server for monitoring purposes, which for security it is not enabled by default.
You can use the monitoring_address configuration option to enable it.
Then just browse the address and human-friendly links to various monitoring and debugging tools should appear.
These include:
- Command-line flags.
- Traces of both short and long lived requests.
- State of the queue.
- State of goroutines.
- Exported variables for whitebox monitoring.
- Profiling endpoints, for use with
go tool pprofor similar tools.
Variables
chasquid exports some variables for monitoring, via the standard expvar package and the OpenMetrics text format, which can be useful for whitebox monitoring.
They're accessible on the monitoring HTTP server, at /debug/vars (default
endpoint for expvars) and /metrics (common endpoint for openmetrics).
The /metrics endpoint is also compatible with
Prometheus.
Note these are still subject to change, although breaking changes will be avoided whenever possible, and will be noted in the release notes.
List of exported variables:
- chasquid/aliases/hookResults (hook result -> counter)
count of aliases hook results, by hook and result. - chasquid/queue/deliverAttempts (recipient type -> counter)
attempts to deliver mail, by recipient type (pipe/local email/remote email). - chasquid/queue/dsnQueued (counter)
count of DSNs that we generated (queued). - chasquid/queue/itemsWritten (counter)
count of items the queue wrote to disk. - chasquid/queue/putCount (counter)
number of envelopes put in the queue. - chasquid/smtpIn/acceptErrors (counter)
count of errors accepting incoming connections. - chasquid/smtpIn/activeConnections (gauge)
number of active incoming connections. - chasquid/smtpIn/commandCount (map of command -> count)
count of SMTP commands received, by command. Note that for unknown commands we useunknown<COMMAND>. - chasquid/smtpIn/connectionsOverLimit (counter)
count of incoming connections closed due to the connection limit, which is half of the file descriptor limit. - chasquid/smtpIn/dkimSignErrors (counter)
count of DKIM sign errors - chasquid/smtpIn/dkimSigned (counter)
count of successful DKIM signs - chasquid/smtpIn/dkimVerifyErrors (counter)
count of DKIM verification errors - chasquid/smtpIn/dkimVerifyFound (counter)
count of messages with at least one DKIM signature - chasquid/smtpIn/dkimVerifyNotFound (counter)
count of messages with no DKIM signatures - chasquid/smtpIn/dkimVerifyValid (counter)
count of messages with at least one valid DKIM signature - chasquid/smtpIn/hookResults (result -> counter)
count of hook invocations, by result. - chasquid/smtpIn/loopsDetected (counter)
count of email loops detected. - chasquid/smtpIn/responseCodeCount (code -> counter)
count of response codes returned to incoming SMTP connections, by result code. - chasquid/smtpIn/securityLevelChecks (result -> counter)
count of security level checks on incoming connections, by result. - chasquid/smtpIn/spfResultCount (result -> counter)
count of SPF checks, by result. - chasquid/smtpIn/tlsCount (tls status -> counter)
count of TLS statuses (plain/tls) for incoming SMTP connections. - chasquid/smtpIn/wrongProtoCount (command -> counter)
count of commands for other protocols (e.g. HTTP commands). - chasquid/smtpOut/securityLevelChecks (result -> counter)
count of security level checks on outgoing connections, by result. - chasquid/smtpOut/sts/mode (mode -> counter)
count of STS checks on outgoing connections, by mode (enforce/testing). - chasquid/smtpOut/sts/security (result -> counter)
count of STS security checks on outgoing connections, by result (pass/fail). - chasquid/smtpOut/tlsCount (status -> counter)
count of TLS status (insecure TLS/secure TLS/plain) on outgoing connections. - chasquid/sourceDateStr (string)
timestamp when the binary was built, in human readable format. - chasquid/sourceDateTimestamp (int)
timestamp when the binary was built, in seconds since epoch. - chasquid/sts/cache/expired (counter)
count of expired entries in the STS cache. - chasquid/sts/cache/failedFetch (counter)
count of failed fetches in the STS cache. - chasquid/sts/cache/fetches (counter)
count of total fetches in the STS cache. - chasquid/sts/cache/hits (counter)
count of hits in the STS cache. - chasquid/sts/cache/invalid (counter)
count of invalid policies in the STS cache. - chasquid/sts/cache/ioErrors (counter)
count of I/O errors when maintaining the STS cache. - chasquid/sts/cache/marshalErrors (counter)
count of marshaling errors when maintaining the STS cache. - chasquid/sts/cache/refreshCycles (counter)
count of STS cache refresh cycles. - chasquid/sts/cache/refreshErrors (counter)
count of STS cache refresh errors. - chasquid/sts/cache/refreshes (counter)
count of STS cache refreshes. - chasquid/sts/cache/unmarshalErrors (counter)
count of unmarshaling errors in the STS cache. - chasquid/version (string)
version string.