Error Trend Chart Generator

Error counts skip around; spotting whether failures are accelerating needs a chart, not eyeballs. Error Trend Chart Generator detects error-level lines and HTTP 4xx/5xx from any pasted log, buckets them by the interval you pick, and draws a proportional bar chart of failures over time using block characters — readable in a plain-text report, no image or JS charting library required.

Consecutive-bucket deltas (▲ rising, ▼ falling, = flat) mark where the failure count rose or fell from one interval to the next, a rising-bucket stat flags sustained growth across the whole window, and the repeated-error ranking names the loudest recurring failure text — an escalation headline in one screenshot.

A line is treated as an error if it contains a recognized error token (ERROR, FATAL, PANIC, EXCEPTION, TRACEBACK, FAILED, DENIED, TIMEOUT) or an HTTP status code in the 400–599 range, whether that code sits mid-line or is the very last token. Anything else — INFO/DEBUG lines, 2xx/3xx statuses — is ignored so the chart reflects failures only, not overall traffic volume.

• Error detection — recognized error-level tokens plus HTTP 4xx/5xx status codes, anywhere in the line.
• Time buckets — 1 minute, 5 minutes, 15 minutes, 1 hour or 1 day, selectable per run.
• Proportional bar chart — block-character bars sized relative to the peak bucket, rendered as plain text.
• Delta markers — ▲/▼/= next to each bucket shows the change from the one before it.
• Rising-bucket stat — a single number counting how many buckets increased over their predecessor, a quick read on whether things are getting worse.
• Top error messages — the most repeated error text (with timestamps stripped) ranked by frequency, so you know which failure is actually recurring.
• Local only — nothing uploaded; charting happens entirely in your browser.

1. Paste log lines that include timestamps at the start of each line — mixed error levels and HTTP status lines are both fine.
2. Pick a bucket interval that matches the time span you pasted: minutes for a short incident window, hours or days for a longer trend.
3. Chart. The tool counts errors per bucket, draws proportional bars, marks the delta against the previous bucket, and ranks the most repeated error messages.
4. Check the rising-buckets stat to see at a glance whether the trend is accelerating or settling down.
5. Copy the full chart report — bars, deltas and top messages — to paste into an incident channel or postmortem.

Example 1 — Escalation. Three consecutive ▲ buckets say the failure rate is compounding, not just noisy — actionable in seconds instead of after staring at a raw log for minutes.

Example 2 — Resolution proof. After deploying a fix, consecutive ▼ buckets in the following minutes confirm the error signal actually dropped, rather than just assuming the fix worked.

Example 3 — Recurrence. The top-message ranking shows the exact same exception text appearing in bucket after bucket — that is a known, unaddressed bug, not random noise, and it deserves a ticket.

Example 4 — Interval zoom. Start with a 1-minute bucket to find the exact moment an incident began, then re-chart the same paste at 1-hour buckets to see how it fits into the day’s overall pattern.

• Failure-focused — successful requests and non-error log lines are excluded, so the chart is signal, not traffic volume.
• Direction aware — delta markers show momentum, not just raw counts, so a small but growing trend doesn’t get lost next to a large but shrinking one.
• Interval zoom — the same paste can be viewed at minute or day granularity without re-collecting data.
• Culprit ranking — the recurring error text is surfaced automatically instead of requiring a manual grep.
• Private — all processing happens client-side; logs never leave the page.