Your CloudWatch logs, diagnosed locally.

Instant insights on _

No queries, no uploads, no data sent anywhere.

Start Analyzing Free

46 patterns · <100ms analysis · 100% browser-local

Interactive Demo

Explore a real analysis.

Scroll through actual results from a Lambda with DynamoDB throttling issues.

Lambda · DynamoDB Throttling

Lambda

payment-processor

Feb 23 14:30 – 15:15 UTC

142 entries

Save New Analysis

Error Rate

23.8%

34 of 142

Cold Starts

18.3%

avg 890ms init

Duration P99

12.40s

avg 1.8s

Memory Peak

87%

of 512MB

Analysis Summary

"error_rate": 23.8,
"cold_start_pct": 18.3,
"duration_p99": 12400,
"memory_peak_pct": 87,
"total_invocations": 142,
"service": "payment-processor"

Findings · 4

CRITICAL · 18

Elevated Error Rate

23.8% of invocations are failing - significantly above the 1% threshold.

→ Add error handling and retry logic for transient failures.

HIGH · 6

Cold Start Latency

18.3% cold start rate with avg 890ms init. P99 heavily impacted.

→ Enable provisioned concurrency or optimize package size.

ELEVATED · 3

Timeout Pattern

3 invocations timed out at the 30s limit.

MEDIUM

Memory Pressure

Peak memory at 87% of 512MB allocation.

Root Cause Analysis · 3

CRITICAL correlation

Error spike correlates with DynamoDB throttling events - 94% of failures occur within 200ms of a ProvisionedThroughputExceededException.

a7f3e2d1 b8c4f1a2 c9d5e0b3 +15 more

▸3 sample(s)

ProvisionedThroughputExceededException: Rate exceeded for table Payments
DynamoDB request failed after 3 retries - backing off 1200ms
Task timed out after 30.00 seconds

→ Switch Payments table to on-demand capacity or increase provisioned WCU.

HIGH cold-start

Cold start init duration (890ms avg) dominated by SDK client initialization. Warm invocations average 45ms.

▸2 sample(s)

INIT_START Runtime Version: nodejs20.x
Duration: 892.41 ms  Init Duration: 891.33 ms

→ Move SDK client init outside the handler. Enable provisioned concurrency.

ELEVATED anomaly

Memory usage trend increasing - 3 recent invocations above 95% allocation, suggesting a slow memory leak.

→ Profile heap usage. Check for accumulating closures or cached objects.

Log Signatures · 4 unique

48x 12 requests error

ProvisionedThroughputExceededException: Rate exceeded for table *

▸show sample

ProvisionedThroughputExceededException: Rate exceeded for table Payments (Service: AmazonDynamoDBv2; Status Code: 400)

142x 142 requests

START RequestId: * Version: $LATEST

26x 26 requests

INIT_START Runtime Version: nodejs20.x

8x 3 requests error

Task timed out after * seconds

Try it on your logs

Root Cause Analysis

Stop guessing.
Start fixing.

Correlates error patterns, memory pressure, and duration spikes into ranked root-cause findings. You see what to fix - not just what broke.

P50/P99/P99.9 duration breakdowns
Memory usage vs. allocation tracking
Timeout, throttle, and OOM detection
Findings ranked CRITICAL → LOW

Root Cause Analysis

3 findings

CRITICAL · 12

Memory pressure causing OOM risk

Avg memory usage is 94% of allocation. 12 invocations near the limit.

→ Increase memory allocation from 512MB to 1024MB

HIGH · 8

Cold starts correlating with P99 spikes

18% cold start rate. P99 duration 3.2s vs P50 of 180ms.

→ Enable provisioned concurrency for this function

ELEVATED · 3

Recurring timeout pattern detected

3 of 47 invocations timed out at 6.4% error rate.

→ Review downstream service latency and timeout config

Browser Extension

Analyze directly from
the CloudWatch console.

Skip the export step entirely. Our browser extension adds a one-click analyze button to the CloudWatch Logs console - select a log group, click analyze, and get findings without ever leaving AWS.