Skip to main content
Card class: HeroCategory: Marketplace
Triggered when /upload/listings has no confirmed ack in 48h, feed is blind, all listing changes stuck.

At a glance

P1 alert. Fires when the Alibris outbound listings feed (/upload/listings) has not received a confirmed ack from Alibris in 48 hours. While active, every listing edit is stuck in a queue with no path to Alibris. The catalogue is frozen on the last successful cycle’s state.
What it countsBoolean alert state, 1 if time_since_last_acked_outbox_upload > 48h, else 0. Maintains a list of queued-but-undelivered outbox files.
API endpoint + reportDerived from the Pure-FTPd connector’s last-successful-ack heartbeat. Vortex IQ polls Alibris’s /download/acks directory every 4 hours.
SeverityP1. Every hour the alert is open, the merchant is operating blind on Alibris. Frozen catalogue; new listings invisible; sold-elsewhere stock still orderable on Alibris (stockout cancellations); price changes don’t propagate.
Fees / commissionNot applicable directly. Indirect: stockout cancellations during outage count toward seller-error cancellation rate (3% threshold).
Refunds / cancellationsNot directly. Downstream: cancellation spike during and 7 days after; watch Cancellation Rate.
CurrencyNot applicable.
Common root causes(1) Pure-FTPd credential rotation not reflected in connector config (~32%); (2) Alibris-side maintenance window overran (~22%); (3) Network connectivity loss (~18%); (4) Outbound file format violation silently rejected (~15%); (5) Disk-full or quota on outbox staging (~9%); (6) Other (~4%).
Multi-marketplace overlapEach marketplace runs its own feed. Co-firing with AbeBooks Feed Cycle Failure suggests merchant-side cause.
Time windowRT.
Alert triggerno successful upload >48h.
Rolesowner, operations, engineering.

Calculation

Calculated automatically from your Alibris data. See the At a glance summary above for what the metric tracks and the worked example below for a typical reading.

Worked example

A US bookseller, 32,100 active Alibris listings, 4-hourly outbound feed cycle. Incident timeline 28 Apr 26 to 01 May 26.
Timestamp (UTC)EventOperational consequence
28 Apr 04:00Last successful outbox upload + ackCatalogue snapshot frozen
28 Apr 04:00 to 30 Apr 04:004-hourly upload cycle attempts continue, no ackFiles queued; no visible failure to merchant
30 Apr 04:00Alert threshold crossed (48h since last ack)Card fires
30 Apr 09:14Ops team paged-
30 Apr 11:32Root cause: Pure-FTPd password rotated 27 Apr; connector still using oldStockout cancellations: 9 between 28 Apr and 30 Apr
30 Apr 12:48Connector config updated; reconnect succeedsFirst successful upload acked
30 Apr 13:14Backlog of 14 queued cycles re-uploaded and ackedAll catalogue changes propagate
30 Apr 14:02Card clears; cancellation-rate effect persists 7 days-
Cumulative impact during 56-hour outage:
Impact dimensionValue
Listings frozen32,100
Stock changes queued~340
Stockout cancellations9
Repricer updates queued~9,200
Estimated revenue impact880to880 to 1,400
Cancellation rate impact (30D rolling)+1.4 percentage points
Six things to notice that are specific to Alibris and the Pure-FTPd feed model:
  1. 48 hours is the standard threshold but tune to your cycle. 4-hourly cycle means a 12 or 24h alert threshold is more appropriate.
  2. Alibris’s Pure-FTPd credentials don’t rotate on a schedule but partner-side network drops happen. Subscribe to Alibris developer notifications.
  3. Stockout cancellation cascade is the second-order damage. During the 56h outage, 340 books sold elsewhere were still listed as available on Alibris.
  4. The 30 to 90 day reputational tail. A 9-cancellation cluster sits in the 30D average for the next 30 days; expect “at risk” seller-standing tile through early Jun 26.
  5. The 9,200 queued repricer updates. Letting them all fire post-outage causes a one-day price-change spike; phased re-upload (1,000/hour) is gentler.
  6. Library Services / institutional buyers may be MORE forgiving than retail. Net-30 procurement cycles mean the institutional cohort doesn’t notice a 2-day outage; retail buyers do.

Sibling cards merchants should reference together

Feed cycle failure is the P1 operational alert. Pair with these:
CardWhy pair it with Feed Cycle Failure
Last Successful UploadCompanion card; shows hours-since-last-good-cycle.
FTP Connection / Auth HealthAuth failures precede feed failures by 4 to 24h.
Failed Batches (7d)7-day history view.
Top Upload Error TypesCause attribution post-recovery.
Inbound Orders File LagCross-direction check.
Cancellation RateDownstream impact.
Suspended ListingsDownstream impact.
AbeBooks Feed Cycle FailureCo-fire confirms merchant-side cause.

Reconciling against the vendor’s own dashboard

Where to look in the Alibris seller dashboard: Alibris’s seller dashboard does NOT publish a feed-health view. Three indirect views help during an active alert:
  1. Sellers → Inventory → Upload History. Last N upload attempts; if most recent is >48h old, Alibris confirms it has not received your files.
  2. Sellers → Inventory → Manage. Spot-check a recently-edited listing; if Alibris shows OLD state, change hasn’t propagated.
  3. Alibris Developer Status Page (intermittent service).
Why our alert may differ from Alibris’s observations:
ReasonDirectionWhy
Alibris-side ack delayCard alerts but feed is fine99th-percentile delays can trip false alerts. Wait one cycle.
Connector polling cadenceCard alerts 4h late4-hour poll cycle.
Outbox-empty edge caseCard silent during quiet period72h with no listing changes won’t trigger.
Manual upload via Alibris UICard unawareManual CSV uploads bypass Pure-FTPd.
Cross-connector reconciliation:
CardExpected relationshipWhat causes legitimate divergence
abebooks.ab_alert_feed_failureIndependent feeds.Co-fires suggest merchant-side cause.
amazon.amzn_alert_feed_failureDifferent protocol (SP-API).Different failure modes.
shopify.alert_webhook_failureDifferent mechanic.Webhook-push vs FTP-poll.

Known limitations / merchant FAQs

The alert just fired. Next 30 minutes? (1) Test Pure-FTPd auth manually with stored credentials. (2) Check outbox staging directory for queued files >48h old. (3) Open Alibris developer status page. Alert cleared but cancellation rate still climbing. Why? Operational damage takes 7 to 30 days to age out of the rolling 30D window. Stockout cancellations during outage stay in the rolling average. Multi-marketplace, can a single network outage trip both Alibris and AbeBooks alerts? Yes if cause is at merchant data centre. Marketplace-side outages affect only one. Listing-quality / Buy Box, second-order damage? Cancellation cascade is primary. Beyond that, Listing Quality Score sees no positive signals during outage and decays slightly. Inventory-sync lag during outage, what’s the right behaviour? (a) Conservative: manually mark all Alibris listings out-of-stock. (b) Permissive: leave live, accept stockout cancellations. Most under-1,000-orders/day sellers use (b). ISBN match quality, can it cause feed failures? Indirectly. Severely malformed ISBN can cause Alibris to silently reject the entire file. Why doesn’t Alibris email me when my feed breaks? Alibris does send weekly batch emails; cadence is 7+ days. This card’s 48h threshold catches issues much faster. Rare books vs commodity, outage impact difference? Commodity bears most operational damage (high volume = many stockout cancellations). Rare books rarely produce same-day stockouts during 56h outages. When does the alert turn off vs when do I act? Alert auto-clears on next ack. But “act” until you’ve drained the outbox queue, confirmed listing states via spot-check, and proactively contacted affected buyers from the outage. Alibris-specific: Library Services / institutional cohort, more or less affected by outages? Less. Net-30 procurement cycles mean institutional buyers don’t browse daily; a 2 to 3 day outage is invisible to them. Retail buyers do browse daily; they bear most reputational damage.

Tracked live in Vortex IQ Nerve Centre

Feed Cycle Failure is one of hundreds of KPI pulses Vortex IQ tracks across Alibris and 70+ other ecommerce connectors. Nerve Centre runs the detection layer; Vortex Mind investigates the cause when something moves; Ask Viq lets you interrogate any number in plain English. Start for free or book a demo to see this metric running on your own data.