Skip to main content
Telebot surfaces errors from handlers, the Telegram API, and the bot internals through a unified pipeline. This page explains each layer and how to configure them.

​
How Handler Errors Flow

When a handler returns a non-nil error, Telebot passes it to the OnError callback that was configured in Settings. If OnError is not set, the default implementation logs the error with the update ID: 
// From bot.go — the default fallback
var defaultOnError = func(err error, c Context) {
    if c != nil {
        log.Println(c.Update().ID, err)
    } else {
        log.Println(err)
    }
}
The bot does not crash on handler errors. Each update is processed independently and errors are reported, then discarded.

​
Configuring a Global Error Handler

Settings.OnError is a function with the signature func(error, Context). The Context parameter may be nil when the error originates outside a handler (for example, during polling). 
// Settings.OnError is a callback function that will get called on errors
// resulted from the handler. It is used as post-middleware function.
// Notice that context can be nil.
OnError func(error, Context)

b, err := tele.NewBot(tele.Settings{
    Token:  os.Getenv("TOKEN"),
    Poller: &tele.LongPoller{Timeout: 10 * time.Second},
    OnError: func(err error, c tele.Context) {
        if c != nil {
            log.Printf("handler error for update %d: %v", c.Update().ID, err)
        } else {
            log.Printf("bot error: %v", err)
        }
        // Optionally report to an external service
        sentry.CaptureException(err)
    },
})
You can also call the error handler manually at any time: 
// OnError triggers the configured error callback.
func (b *Bot) OnError(err error, c Context)

​
Telebot Error Variables

The telebot package defines sentinel errors for common bot-level failure cases: 
var (
    ErrBadRecipient    = errors.New("telebot: recipient is nil")
    ErrUnsupportedWhat = errors.New("telebot: unsupported what argument")
    ErrCouldNotUpdate  = errors.New("telebot: could not fetch new updates")
    ErrTrueResult      = errors.New("telebot: result is True")
    ErrBadContext      = errors.New("telebot: context does not contain message")
)
ErrorWhen it occurs
ErrBadRecipientb.Send was called with a nil recipient
ErrUnsupportedWhatThe what argument passed to Send/Reply is not a string or Sendable
ErrCouldNotUpdateThe long poller could not fetch updates from Telegram
ErrTrueResultTelegram responded with {"ok":true,"result":true} instead of an object
ErrBadContextA context method that requires a message was called on a non-message update

​
Telegram API Error Types

Errors returned from the Telegram API are represented as *tele.Error: 
type Error struct {
    Code        int
    Description string
    Message     string
}

// Error implements the error interface.
func (err *Error) Error() string
The Error() method formats as "telegram: <message> (<code>)". Two specialised subtypes carry extra information: 
type FloodError struct {
    err        *Error
    RetryAfter int  // seconds to wait before retrying
}

type GroupError struct {
    err        *Error
    MigratedTo int64  // new supergroup chat ID
}

​
Matching API errors

Use errors.Is or the helper tele.ErrIs to match against one of the pre-declared error variables: 
_, err := b.Send(recipient, "Hello")
if errors.Is(err, tele.ErrBlockedByUser) {
    // remove user from the database
    db.DeleteUser(recipient.ID)
    return
}
Or use a type assertion to inspect the raw fields: 
var apiErr *tele.Error
if errors.As(err, &apiErr) {
    log.Printf("telegram error %d: %s", apiErr.Code, apiErr.Description)
}

​
Full list of pre-declared errors

General (4xx/5xx)

ErrTooLarge, ErrUnauthorized, ErrNotFound, ErrInternal

Forbidden (403)

ErrBlockedByUser, ErrKickedFromGroup, ErrKickedFromSuperGroup, ErrKickedFromChannel, ErrNotStartedByUser, ErrUserIsDeactivated, ErrNotChannelMember

Message errors

ErrCantEditMessage, ErrMessageNotModified, ErrSameMessageContent, ErrEmptyMessage, ErrEmptyText, ErrTooLongMessage, ErrTooLongMarkup

File errors

ErrWrongFileID, ErrWrongFileIDCharacter, ErrWrongFileIDLength, ErrWrongFileIDPadding, ErrWrongFileIDSymbol, ErrCantUploadFile, ErrBadURLContent

Rights errors

ErrNoRightsToDelete, ErrNoRightsToRestrict, ErrNoRightsToSend, ErrNoRightsToSendGifs, ErrNoRightsToSendPhoto, ErrNoRightsToSendStickers, ErrCantRemoveOwner, ErrUserIsAdmin

Chat errors

ErrChatNotFound, ErrEmptyChatID, ErrGroupMigrated, ErrChatAboutNotModified, ErrChannelsTooMuch, ErrChannelsTooMuchUser

​
Handling flood errors

var floodErr tele.FloodError
if errors.As(err, &floodErr) {
    log.Printf("rate limited, retrying in %d seconds", floodErr.RetryAfter)
    time.Sleep(time.Duration(floodErr.RetryAfter) * time.Second)
    // retry the operation
}

​
Handling group migration

var groupErr tele.GroupError
if errors.As(err, &groupErr) {
    log.Printf("group migrated to supergroup: %d", groupErr.MigratedTo)
    db.UpdateChatID(oldChatID, groupErr.MigratedTo)
}

​
Error Handling in Middleware

Middleware can inspect, transform, or suppress errors returned by handlers: 
func ErrorMapper(next tele.HandlerFunc) tele.HandlerFunc {
    return func(c tele.Context) error {
        err := next(c)
        if err == nil {
            return nil
        }
        // Suppress "message not modified" — this often happens with
        // idempotent edits and is not worth surfacing
        if errors.Is(err, tele.ErrMessageNotModified) {
            return nil
        }
        return err
    }
}

b.Use(ErrorMapper)

​
Panic Recovery with Recover Middleware

Handlers that panic will crash the goroutine processing the update, taking down that update’s processing. The middleware.Recover middleware wraps every handler in a deferred recover: 
// Recover returns a middleware that recovers a panic happened in
// the handler.
func Recover(onError ...RecoverFunc) tele.MiddlewareFunc
It captures both error-typed panics and string-typed panics, converting them to errors that are then routed through b.OnError (or a custom RecoverFunc). 
b.Use(middleware.Recover())
Register Recover as the first middleware (outermost) with b.Use so it wraps the entire middleware chain. Middleware registered before Recover won’t be protected.
// Recommended ordering
b.Use(middleware.Recover()) // outermost — catches panics from everything below
b.Use(middleware.Logger())  // logs all requests
b.Use(AuthMiddleware)       // custom auth

​
Best Practices for Production Bots

Always set OnError

The default handler only logs to stderr. In production, route errors to a structured logger or an error-tracking service (Sentry, Datadog, etc.).

Use Recover middleware

Panics in handlers should never crash the bot. Always register middleware.Recover() globally, before other middleware.

Match specific API errors

Don’t treat all API errors the same. Check for ErrBlockedByUser to clean up stale users, ErrMessageNotModified to suppress noise, and FloodError to implement backoff.

Return errors from handlers

Always return errors from handler functions instead of swallowing them with log.Println. This ensures they flow through OnError and can be centrally tracked.

​
Structured error handling example

func setupBot() *tele.Bot {
    b, err := tele.NewBot(tele.Settings{
        Token:  os.Getenv("TOKEN"),
        Poller: &tele.LongPoller{Timeout: 10 * time.Second},
        OnError: func(err error, c tele.Context) {
            // Classify and route the error
            var apiErr *tele.Error
            var floodErr tele.FloodError

            switch {
            case errors.Is(err, tele.ErrBlockedByUser),
                errors.Is(err, tele.ErrUserIsDeactivated),
                errors.Is(err, tele.ErrNotStartedByUser):
                // User is unreachable — clean up silently
                if c != nil {
                    db.DeactivateUser(c.Sender().ID)
                }

            case errors.As(err, &floodErr):
                // Telegram is rate limiting us
                log.Printf("flood control: retry after %ds", floodErr.RetryAfter)

            case errors.Is(err, tele.ErrMessageNotModified):
                // Idempotent edit — not worth logging

            case errors.As(err, &apiErr) && apiErr.Code >= 500:
                // Telegram server error — report and alert on-call
                sentry.CaptureException(err)

            default:
                // Unexpected error — always report
                sentry.CaptureException(err)
                if c != nil {
                    log.Printf("unhandled error (update %d): %v", c.Update().ID, err)
                }
            }
        },
    })
    if err != nil {
        log.Fatal(err)
    }

    // Protect against panics
    b.Use(middleware.Recover())

    return b
}

Build docs developers (and LLMs) love

Get started for freeTalk to us