Error handling

Learn how to handle errors when using Redis client libraries.

When working with Redis, errors can occur for various reasons, including network issues, invalid commands, or resource constraints. This guide explains the types of errors you might encounter and how to handle them effectively.

Categories of errors

Redis errors fall into four main categories. The table below provides a quick overview of each type. Click on any error type to jump to its detailed section, which includes common causes, examples, handling strategies, and code examples.

Connection errors

Connection errors occur when your application cannot communicate with Redis. These are typically temporary and often recoverable.

Common causes:

• Network connectivity issues
• Redis server is down or unreachable
• Authentication failure
• Connection timeout
• Connection pool exhaustion

Examples:

• ConnectionError: Network failure or server unreachable
• TimeoutError: Operation exceeded the configured timeout
• AuthenticationError: Invalid credentials

When to handle: Almost always. Connection errors are usually temporary, so implementing retry logic or fallback strategies is recommended.

Example strategy:

    graph TB
    A["Try to connect<br/>to Redis"]
    A -->|Success| B(["Use the result"])
    A -->|Failure| C{Error type?}
    C -->|Timeout| D(["Retry with<br/>exponential backoff"])
    C -->|Auth failure| E(["Check credentials<br/>and fail"])
    C -->|Network error| F(["Fallback to<br/>alternative data source"])

Command errors

Command errors occur when Redis receives an invalid or malformed command. These typically indicate a bug in your code.

Common causes:

• Typo in command name
• Wrong number of arguments
• Invalid argument types (for example, supplying a string key to a list command))
• Using a command that doesn't exist in your Redis version

Examples:

• ResponseError: Invalid command or syntax error
• WRONGTYPE Operation against a key holding the wrong kind of value
• ERR unknown command

When to handle: Rarely. These usually indicate programming error and so you should fix the errors in your code rather than attempt to handle them at runtime. However, some cases (like invalid user input) may be worth handling.

Example:

    graph TB
    A["User provides<br/>JSONPath expression"]
    A --> B["Try to execute it"]
    direction TB
    B -->|ResponseError| C["Log the error"]
    C --> D(["Return default value<br/>or error message to user"])
    B -->|Success| E(["Use the result"])

Data errors

Data errors occur when there are problems with the data itself, such as serialization failures, or data corruption.

Common causes:

• Object cannot be serialized to JSON
• Cached data is corrupted
• Attempting to deserialize invalid data

Examples:

• JSONDecodeError: Cannot deserialize JSON data
• SerializationError: Cannot serialize object

When to handle: Sometimes. If the error is due to user input or external data, handle it gracefully. If it's due to your code, fix the code.

Example:

    graph TB
    A["Read cached data"]
    A --> B["Try to deserialize"]
    B -->|Success| C(["Use the data"])
    B -->|Deserialization fails| D["Log the error"]
    D --> E["Delete corrupted<br/>cache entry"]
    E --> F(["Fetch fresh data<br/>from source"])

Resource errors

Resource errors occur when Redis runs out of resources or hits limits.

Common causes:

• Memory limit reached
• Connection pool exhausted
• Too many connections
• Key eviction due to memory pressure

Examples:

• OOM command not allowed when used memory > 'maxmemory'
• Connection pool timeout
• LOADING Redis is loading the dataset in memory

When to handle: Sometimes. Some resource errors are temporary (Redis loading), while others indicate a configuration problem.

Example:

    graph TB
    A{Resource error<br/>occurred?}
    A -->|Redis loading| B(["Retry after<br/>a delay"])
    A -->|Memory full| C(["Check Redis<br/>configuration and data"])
    A -->|Pool exhausted| D(["Increase pool size<br/>or reduce concurrency"])

Error handling patterns

Pattern 1: Fail fast

Use this when the error is unrecoverable or indicates a bug in your code.

When to use:

• Command errors (invalid syntax)
• Authentication errors
• Programming errors

Example:

try:
    result = r.get(key)
except redis.ResponseError as e:
    # This indicates a bug in our code
    raise  # Re-raise the exception

Pattern 2: Graceful degradation

Use this when you have an alternative way to get the data you need, so you can fall back to using the alternative instead of the preferred code.

When to use:

• Cache reads (fallback to database)
• Session reads (fallback to default values)
• Optional data (skip if unavailable)

Example:

try:
    cached_value = r.get(key)
    if cached_value:
        return cached_value
except redis.ConnectionError:
    logger.warning("Cache unavailable, using database")

# Fallback to database
return database.get(key)

Pattern 3: Retry with backoff

Use this when the error could be due to network load or other temporary conditions.

When to use:

• Connection timeouts
• Temporary network issues
• Redis loading data

Example:

import time

max_retries = 3
retry_delay = 0.1

for attempt in range(max_retries):
    try:
        return r.get(key)
    except redis.TimeoutError:
        if attempt < max_retries - 1:
            time.sleep(retry_delay)
            retry_delay *= 2  # Exponential backoff
        else:
            raise

Note that client libraries often implement retry logic for you, so you may just need to provide the right configuration rather than implementing retries yourself. See Client-specific error handling below for links to pages that describe retry configuration for each client library.

Pattern 4: Log and continue

Use this when the operation is not critical to your application.

When to use:

• Cache writes (data loss is acceptable)
• Non-critical updates
• Metrics collection

Example:

try:
    r.setex(key, 3600, value)
except redis.ConnectionError:
    logger.warning(f"Failed to cache {key}, continuing without cache")
    # Application continues normally

Decision tree: How to handle errors

    graph LR
    Start{Error occurred?}

    Start -->|Connection error| C1{Operation type?}
    C1 -->|Read| C2["Graceful degradation<br/>fallback"]
    C1 -->|Write| C3["Log and continue<br/>or retry"]
    C1 -->|Critical| C4["Retry with backoff"]

    Start -->|Command error| Cmd1{Error source?}
    Cmd1 -->|User input| Cmd2["Log and return<br/>error to user"]
    Cmd1 -->|Your code| Cmd3["Fail fast<br/>fix the bug"]

    Start -->|Data error| D1{Operation type?}
    D1 -->|Read| D2["Log, invalidate,<br/>fallback"]
    D1 -->|Write| D3["Log and fail<br/>data is invalid"]

    Start -->|Resource error| R1{Error type?}
    R1 -->|Redis loading| R2["Retry with backoff"]
    R1 -->|Pool exhausted| R3["Increase pool size"]
    R1 -->|Memory full| R4["Check configuration"]

Logging and monitoring

In production, you may find it useful to log errors when they occur and monitor the logs for patterns. This can help you identify which errors are most common and whether your retry and fallback strategies are effective.

What to log

• Error type and message: What went wrong?
• Context: Which key? Which operation?
• Timestamp: When did it happen?
• Retry information: Is this a retry? How many attempts?

Example:

logger.error(
    "Redis operation failed",
    extra={
        "error_type": type(e).__name__,
        "operation": "get",
        "key": key,
        "attempt": attempt,
        "timestamp": datetime.now().isoformat(),
    }
)

What to monitor

• Error rate: How many errors per minute?
• Error types: Which errors are most common?
• Recovery success: How many retries succeed?
• Fallback usage: How often do we use fallback strategies?

These metrics help you identify patterns and potential issues.

Common mistakes

Catching all exceptions

Problem: If you catch all exceptions, you might catch unexpected errors and hide bugs.

Example (wrong):

try:
    result = r.get(key)
except Exception:  # Too broad - some errors indicate code problems.
    pass

Better approach: Catch specific exception types.

Example (correct):

try:
    result = r.get(key)
except redis.ConnectionError:
    # Handle connection error
    pass

Not distinguishing error types

Problem: Different errors need different handling. For example, retrying a syntax error won't help.

Example (wrong):

try:
    result = r.get(key)
except redis.ResponseError:
    # Retry? This won't help if it's a syntax error.
    retry()

Better approach: Handle each error type differently based on whether or not it is recoverable.

Example (correct):

try:
    result = r.get(key)
except redis.TimeoutError:
    retry()  # Retry on timeout
except redis.ResponseError:
    raise   # Fail on syntax error

Ignoring connection pool errors

Problem: Connection pool errors indicate a configuration or concurrency issue that needs to be addressed.

Example (wrong):

# Pool is exhausted, but we don't handle it
result = r.get(key)  # Might timeout waiting for connection

Better approach: Monitor pool usage and increase size if needed.

Client-specific error handling

For detailed information about exceptions in your client library, see:

• redis-py error handling
• Node.js error handling
• Java (Jedis) error handling
• Go (go-redis) error handling
• .NET (NRedisStack) error handling