Elixir: (DBConnection.ConnectionError) tcp recv: closed

Quick fix

Read the commands before running them. Anything that restarts a service, deletes data or changes permissions should be tried on a non-production system first.

# Keep pooled sockets younger than the shortest idle timeout on the path
config :my_app, MyApp.Repo,
  pool_size: 10,
  idle_interval: 15_000,        # ping idle connections
  queue_target: 50,
  queue_interval: 1000

# What is actually closing it: check the database side first
# Postgres
SHOW idle_in_transaction_session_timeout;
SELECT state, count(*) FROM pg_stat_activity GROUP BY state;
# pgbouncer: server_idle_timeout, server_lifetime, and pool_mode

# Transaction pooling rejects prepared statements, which looks like random closes
config :my_app, MyApp.Repo, prepare: :unnamed

# Long running queries killed by a server timeout give the same message
config :my_app, MyApp.Repo, timeout: 15_000

# Retrying is correct here: the next checkout gets a fresh socket

How to diagnose Elixir errors

Elixir errors are shaped by the actor model: a GenServer.call timeout does not mean the server crashed, it means the server was busy for longer than the caller was willing to wait. The right question is usually "what is that process doing?" rather than "why did the call fail?". Because supervisors restart failed processes automatically, transient errors can also hide in the logs while the system appears healthy.

If the quick fix above does not resolve it, work through these steps. They apply to this whole class of error, not just to this one message, which is usually what saves the time.

1. Attach to a running node with iex --remsh and inspect the process: Process.info(pid, :message_queue_len). A growing mailbox means the process is the bottleneck.
2. Use :observer.start() to see the supervision tree, process memory and message queues visually.
3. Move long-running work out of handle_call. Use handle_cast, a Task, or a dedicated pool so the GenServer stays responsive.
4. Check restart intensity. A supervisor that exceeds max_restarts takes down its own supervisor, producing a cascade that looks like an unrelated failure at the top.
5. For compile errors in dependencies, run mix deps.compile --force and check that any required native toolchain (make, gcc, erlang headers) is installed.

Tools worth reaching for

• :observer.start()
• iex --remsh
• Process.info/2
• mix deps.tree
• :recon

Authoritative references

Primary documentation for this error, worth reading before applying any fix in production.

📖 hexdocs.pm 📖 hexdocs.pm

Related Elixir errors

• Ecto: connection not available and request was dropped from queueEvery connection in the pool was busy for longer than queue_target allowed, so DBConnection…
• Elixir: (FunctionClauseError) no function clause matchingEvery clause of the function was tried and none matched the arguments, either on the pattern…
• Elixir: (KeyError) key :name not found in: %{"name" => "ada"}Dot access and `Map.fetch!/2` raise rather than returning nil when a key is missing, and the…
• Elixir: (MatchError) no match of right hand side value: {:error, :enoent}A pattern match assumed the happy path and the function returned the error tuple instead, so…
• Elixir: function Foo.bar/1 is undefined (module Foo is not available)The module was never loaded, which in practice means a misspelled alias, a file outside the…
• Elixir: GenServer call timeout (exited in :gen_server.call)A GenServer.call did not get a reply within 5000ms (default) because the server is…
• Elixir: Mix could not compile dependencyA dependency failed to compile due to missing native build tools, a version conflict, or…
• Elixir: protocol Enumerable not implemented for nil of type AtomSomething handed `Enum` a nil where a list was expected, and nil has no Enumerable…

See all 12 Elixir errors →

Browse other categories

• HTTP 494xx client errors, 5xx server errors, redirects, headers and protocol problems.
• JavaScript 42npm resolution, async pitfalls, hydration, memory limits and runtime type…
• Database 41Connections, deadlocks, constraints, replication and memory limits.
• AI 35Rate limits, context windows, GPU memory and model-serving failures.
• Network 35Refused connections, timeouts, resets, MTU problems and port exhaustion.
• Python 35Imports, virtual environments, encoding, concurrency and dependency conflicts.
• Kubernetes 34CrashLoopBackOff, ImagePullBackOff, OOMKilled, RBAC, scheduling and storage.
• Docker 27Daemon connectivity, disk space, image pulls, ports and architecture mismatches.
• System 26Disk space, systemd units, file descriptors, OOM killer and scheduled jobs.
• Cloud 25IAM permissions, quotas, service limits and credential failures.
• Security 25JWT validation, CSRF, OAuth grants, SELinux, SSH host keys and CSP.
• TLS 24Untrusted authorities, expiry, hostname mismatch, chains and cipher negotiation.

Something missing or wrong?

This entry is maintained by hand. If the fix is out of date, incomplete, or you have a better one, email a correction and it will be reviewed.