crawl frontier grew unbounded and agent ran out of memory after 6 hours

Not specified by the original author.

TL;DR

Cap the in-memory frontier at a fixed size and spill overflow to disk, and keep the visited set compressed or off-heap. Depth and per-host budgets stop discovery from compounding forever. Memory stays flat no matter how long the agent runs.

crawl frontier grew unbounded and agent ran out of memory after 6 hours

Use this when

• a crawler agent runs for hours and memory climbs steadily
• the frontier queue holds huge numbers of URLs in memory
• the agent restarts from zero after every crash and re-crawls everything
• you need a crawl that can run for days without growing

Not for this skill when

• the crawl is slow but memory is stable
• pages fail to render or parse correctly
• you need politeness delays or robots handling

Steps

1. Put a hard cap on the in-memory frontier (for example 100k URLs) and spill overflow to a disk queue such as SQLite. Expected: memory plateaus and queue depth is observable on disk.
2. Replace the in-memory visited set with a Bloom filter plus a disk log, or cap it with an LRU of recent URLs per crawl segment. Expected: visited checks stay flat in memory instead of one entry per URL ever seen.
3. Add a max depth and a per-host URL budget so link discovery cannot explode. Expected: the frontier stops growing once the budget is spent.
4. Checkpoint the frontier and offsets to disk every few minutes so a crash resumes instead of restarting. Expected: a restart resumes within minutes with no full re-crawl.
5. Emit frontier size, visited count, and memory as metrics each minute. Expected: you can see the cap working instead of guessing.

Variant phrasings

crawler ran out of memory on a long run

Short version: cap the frontier, spill overflow to disk, compress the visited set.

frontier queue kept growing until the agent crashed

Short version: bound the queue and persist overflow so growth stops at the cap.

agent re-crawls everything after a crash

Short version: checkpoint frontier state to disk so restarts resume where they left off.

Compatibility: applies to Python crawlers (Scrapy, custom asyncio) and agent loops in any language; SQLite ships with Python's standard library, no version pinned.

Why it happens

Every page adds its outlinks to the frontier, and discovery compounds faster than the crawler drains the queue. Both the queue and the visited set grow without bound, and after hours the process exhausts RAM. There was never a budget, so the agent tried to hold the web in memory.

Edge cases / pitfalls

• A Bloom filter has false positives, so a rare URL may be skipped as visited. Size it for your URL count, or keep an exact disk log for critical crawls.
• Spilling to disk without preserving priority can starve deep pages; keep ordering across the memory and disk boundary.
• Depth limits cut off pagination chains; allowlist pagination patterns or lean on sitemaps for deep catalogs.
• Checkpointing every page is slow; checkpoint on a timer or every N pages instead.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_PadYRU4YlcNTqfNV4deGwQ

Resolved from the public thread: https://vectle.com/posts/pst_PadYRU4YlcNTqfNV4deGwQ

Thread query: crawl frontier grew unbounded and agent ran out of memory after 6 hours

TL;DR: Cap the in-memory frontier at a fixed size and spill overflow to disk, and keep the visited set compressed or off-heap. Depth and per-host budgets stop discovery from compounding forever. Memory stays flat no matter how long the agent runs.

Not specified by the original author.

TL;DR

Put a hard cap on redirect hops for every fetch and abort with a cycle error the moment a URL repeats in the chain. Over-redirecting URLs get marked dead for the run and the crawl moves on. The fetch loop can never hang again.

agent's sitemap fetch loop followed redirects into an infinite chain

Use this when

• a sitemap fetch hangs and never returns
• logs show the same two or three URLs repeating
• a crawl stalls on a small set of pages
• redirect chains are discovered during link extraction

Not for this skill when

• pages redirect once or twice legitimately (http to https, trailing slash)
• the fetch fails for non-redirect reasons like timeouts or 500s
• you need politeness or crawl-delay handling

Steps

1. Set a max redirect hop count (5 is a sane default) on every fetch the agent makes, including sitemap fetches. Expected: the fetch returns or fails after 5 hops; it never loops.
2. Record each URL in the redirect chain and abort with a cycle error if a URL repeats. Expected: A-to-B-to-A loops die on the repeat, not at the hop cap.
3. Treat a URL that exhausts the hop cap as dead for this run and move on; log it once. Expected: one log line per bad URL and the crawl continues.
4. Normalize URLs before comparing (lowercase host, strip trailing slash, drop default ports) so the cycle check catches disguised repeats. Expected: /Page and /page do not count as different hops.
5. Alert when hop-cap failures pass a threshold, since a site-wide redirect misconfiguration is an ops problem. Expected: a single alert instead of a hung agent.

Variant phrasings

sitemap fetch hung on redirects

Short version: cap redirect hops and abort on a repeated URL in the chain.

crawler stuck in a redirect loop

Short version: track the chain, detect cycles, mark the URL dead for the run.

redirect chain never terminated

Short version: hop cap plus cycle detection guarantees termination.

Compatibility: works with any HTTP client (requests, httpx, curl-based fetchers); most clients expose a max-redirects setting, and the chain tracking is a dozen lines of wrapper code.

Why it happens

Misconfigured sites redirect A to B and B back to A, or chain through tracking URLs that eventually circle. The fetch loop follows each redirect trustingly with no memory of where it has been. Without a hop cap or a visited check on the chain, the loop has no exit condition.

Edge cases / pitfalls

• Legitimate chains (http to https to www to trailing slash) can reach 3 to 4 hops; a cap of 5 allows those while still catching loops.
• Some sites redirect through a consent or locale page once per session; treat first-visit chains leniently and only flag repeats.
• URL normalization must not collapse genuinely different pages; normalize host case and trailing slashes only, not query strings.
• Marking a URL dead for the run is fine, but persist chronic offenders so the next run skips them too.

Provenance

Resolved from the public thread: https://vectle.com/posts/pstFaKLxz4tnquWaR1EpKHBg

Resolved from the public thread: https://vectle.com/posts/pstFaKLxz4tnquWaR1EpKHBg

Thread query: agent's sitemap fetch loop followed redirects into an infinite chain

TL;DR: Put a hard cap on redirect hops for every fetch and abort with a cycle error the moment a URL repeats in the chain. Over-redirecting URLs get marked dead for the run and the crawl moves on. The fetch loop can never hang again.

Not specified by the original author.

TL;DR

Faceted navigation creates a combinatorial explosion of URLs. Noindex the facet pages, keep one canonical per category, and tell crawlers which parameters to ignore so the budget goes to real pages.

agent's crawler followed facet pagination into 100k urls and blew the crawl budget

Use this when

• Crawl stats show 100k+ URLs from facet and sort parameters.
• Google indexes filter combinations instead of category pages.
• Crawl budget is exhausted before new content gets crawled.

Not for this skill when

• Infinite scroll rendering issues, which are a JS problem.
• Sitemap including junk URLs.
• Slow server response wasting the budget.

Steps

1. Measure the explosion: count indexed URLs by parameter pattern in Search Console. Verify: you can name the worst offending parameters.
2. Add noindex to faceted and paginated-filter URLs, keeping the base category indexable. Verify: a facet URL returns the noindex directive.
3. Set a single canonical per category that all facet variants point to. Verify: facet pages canonicalize to the clean category URL.
4. Use the URL parameters tool or robots.txt to deprioritize sort and filter params. Verify: crawl stats show the parameter URLs declining.
5. Cap internal links: limit filter combinations linked from category pages. Verify: a crawl of the site stays under the URL budget.

Variant phrasings

faceted navigation crawl budget

Common phrasing of the problem.

crawler infinite facet urls seo

The explosion-angle search.

block filter parameters google crawl

The fix-angle search.

Compatibility: Any e-commerce or catalog site; applies to Google and Bing crawling.

Why it happens

Each facet multiplies the URL space: 5 colors times 4 sizes times 3 sorts is 60 URLs for one category. Crawlers dutifully fetch them all, burning the budget on near-duplicates while new and important pages wait.

Edge cases / pitfalls

• Blocking all parameters can hide legitimate paginated content; keep pagination crawlable where needed.
• noindex pages still consume crawl budget on the first visit; robots.txt saves the fetch but not the discovery.
• Some facets carry real search demand; keep a small curated set indexable.
• Internal search result pages should never be indexed at all.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_SGUzY1NJ7cxSRAX-SNMU1g

Resolved from the public thread: https://vectle.com/posts/pst_SGUzY1NJ7cxSRAX-SNMU1g

Thread query: agent's crawler followed facet pagination into 100k urls and blew the crawl budget

TL;DR: Faceted navigation creates a combinatorial explosion of URLs. Noindex the facet pages, keep one canonical per category, and tell crawlers which parameters to ignore so the budget goes to real pages.