Skip to main content

How to troubleshoot high RTB processing time

Diagnose high inbound RTB processing time by checking buyer response latency, timeout settings, fanout, routing complexity, request data, and traffic volume.

Written by Moja Bot

Processing time is the amount of time Moja AI spends handling an inbound RTB request before it can return the bid, no-bid, or routing result. High processing time can delay call routing, reduce connection rates, increase no-bid outcomes, or cause buyers or callers to time out.

The goal is to prove whether the delay is caused by one buyer, too much buyer fanout, timeout settings, routing complexity, invalid request data, buyer capacity, or broader traffic volume/network behavior.

High processing time should be judged relative to your timeout settings, buyer SLA, call routing expectations, and the normal baseline for the same campaign, RTB group, buyer, geography, and time of day. A processing time spike matters most when it correlates with fewer bids, more timeouts, fewer connected calls, failed transfers, or caller abandonment.

Processing time layers

Layer

What it means

Request validation

Moja AI checks required fields, campaign eligibility, and formatting.

Routing/filter evaluation

Moja AI evaluates campaign, routing plan, filters, caps, hours, and duplicate rules.

Buyer request fanout

Moja AI sends requests to eligible buyers or RTB targets.

Buyer response wait

Moja AI waits for bids, rejects, errors, or timeout.

Bid evaluation

Moja AI compares valid bids, payout/economics, and required routing fields.

Routing result

Moja AI selects a buyer/target and returns the destination or no-bid result.

What happens during processing

When an inbound RTB request enters Moja AI, the platform may need to:

  • Validate request data.

  • Check campaign, RTB group, routing, filter, cap, and duplicate rules.

  • Send requests to one or more buyers.

  • Wait for buyer responses until a response or timeout is reached.

  • Evaluate bid amounts, response status, and reject reasons.

  • Select the winning buyer or target when a valid routable result is available.

  • Return the bid, no-bid, or routing result.

Each step adds time. The most common cause of high processing time is waiting on slow buyer responses, but high fanout, long timeouts, complex routing, invalid data, and traffic spikes can also contribute.

Key RTB fields to review

Start in RTB Logs or RTB reporting and compare processing time by buyer, campaign, RTB group, and time period. Look for buyers, geographies, or hours where processing time is consistently higher than normal.

Field

Why it matters

Processing time

Shows total decisioning time for the request.

Buyers attempted

Confirms which buyers or RTB targets were queried.

Buyer response time

Identifies which buyer is slow compared with other buyers.

Buyer response status

Shows bid, no-bid, reject, error, or timeout.

Reject reason

Helps distinguish slow rejection from actual demand loss.

Timeout setting

Shows how long Moja AI waits before moving on or returning a result.

Bid amount

Confirms whether slow buyers are worth waiting for.

Destination / transfer value

Confirms whether the buyer returned a routable result.

Campaign / RTB group

Shows whether slowness is isolated to one route or affects multiple flows.

Request timestamp

Helps spot time-of-day, dayparting, capacity, or volume-related latency.

State / ZIP / custom tags

Helps identify geo-specific, vertical-specific, or data-mapping causes.

How to diagnose high processing time

  1. Pick a time range where processing time was high.

  2. Compare processing time by buyer or target. If one buyer is much slower than the rest, start there.

  3. Compare response status and reject reasons. A slow buyer that returns useful winning bids is different from a slow buyer that mostly times out or rejects.

  4. Check how many buyers were attempted per request. Too much fanout can slow the entire RTB path.

  5. Compare by campaign or RTB group to see whether the issue is isolated to one setup.

  6. Compare by state, ZIP, source, or custom tags to find geo-specific or data-specific patterns.

  7. Compare by hour or day to detect buyer capacity limits, concurrency issues, or traffic spikes.

  8. If high processing time correlates with no-bids, failed transfers, or lower connection rates, review timeout behavior and routing results together.

Slow but useful vs slow and worthless

Do not treat every slow buyer the same. A slow buyer that wins profitable calls should be handled differently from a slow buyer that mostly times out, errors, or no-bids.

Buyer behavior

Recommended action

Slow and rarely bids

Lower priority, shorten timeout, or pause.

Slow but high-value bidder

Keep, but isolate into a separate route or tune timeout carefully.

Fast but low-value bidder

Keep as fallback, but do not let it block better demand.

Slow only during peak hours

Check buyer capacity, concurrency, and dayparting.

Slow with errors/timeouts

Contact buyer or disable until fixed.

Common causes

  • Slow buyer endpoints. Buyer APIs may be overloaded, unstable, rate-limited, or slow to respond.

  • Too many buyers in the RTB flow. More eligible buyers means more requests, responses, statuses, and bids to evaluate.

  • Long timeout settings. If timeouts are too generous, Moja AI may wait too long for buyers that rarely return useful bids.

  • Network or API latency. External buyer services or network paths can introduce delays outside Moja AI.

  • Complex routing logic. Heavy filtering, multiple targets, advanced rules, or broad RTB groups can add processing time.

  • High request volume. Volume spikes can expose buyer capacity limits or make slow routes more visible.

  • Invalid or incomplete data. Invalid or incomplete data can cause rejects, extra rule evaluation, or no-bid outcomes.

  • Geo-specific buyer coverage issues. Some states, ZIPs, or regions may require more buyer evaluation before a usable bid is found.

Troubleshooting table

Signal

Likely cause

What to do

One buyer is slower than all others

Buyer endpoint latency

Contact buyer, lower timeout, deprioritize, or pause.

All buyers are slow

Routing complexity, network issue, or request volume is high

Review campaign/RTB group setup and request volume.

Processing time high but bids still win

Slow high-value buyer

Compare revenue impact before disabling.

Processing time high and no bids

Buyers timing out, rejecting late, or returning no usable demand

Review response status, reject reasons, and timeout behavior.

Processing time spikes at certain hours

Buyer capacity, concurrency, or volume issue

Compare by hour/day and check buyer availability.

Long processing before failed call transfer

Routing result came too late

Shorten timeout and reduce buyer fanout.

Many buyers attempted per request

Too much fanout

Reduce eligible buyers or split routes.

Slow only for certain states/ZIPs

Geo-specific buyer coverage issue

Review filters, coverage, and buyer response by geography.

Invalid data increases processing

Missing or malformed request fields

Validate required fields and custom tag mappings.

How to reduce processing time

Do not lower timeouts blindly. Shorter timeouts can improve speed, but they can also reduce bids or revenue if valuable buyers need more time to respond.

  • Lower timeout settings for buyers that rarely return useful bids.

  • Pause buyers with repeated timeout or error behavior.

  • Deprioritize slow buyers that produce low revenue.

  • Keep high-value slow buyers in a separate path so they do not slow every request.

  • Reduce fanout by filtering buyers before sending requests to them.

  • Confirm required fields are mapped before RTB evaluation.

  • Compare buyer response time against win rate and revenue before removing demand.

  • Watch for time-of-day spikes that indicate buyer capacity limits.

  • Split broad routes into more targeted RTB groups when one path is carrying too many buyer checks.

Related articles

What to send support

For latency issues, support needs exact RTB examples and a time window. Send:

  • RTB request/log ID examples

  • Campaign or RTB group

  • Buyer/target names involved

  • Time range and timezone

  • Processing time observed

  • Normal or baseline processing time, if known

  • Buyer response statuses

  • Timeout settings

  • Whether the issue affects one buyer, one campaign, or all traffic

  • Whether it correlates with no-bids, failed transfers, or lower connection rates

Summary

High processing time usually means Moja AI is waiting on slow buyer responses, evaluating too many buyers, using long timeout settings, handling complex routing, or processing traffic with missing or hard-to-match data. Start by comparing processing time by buyer, campaign/RTB group, time period, response status, and revenue impact before removing demand.

Did this answer your question?