A video playing to completion does not establish that its verification resource executed. Playback and measurement can have different outcomes, and a useful incident record preserves both.
VAST's verificationNotExecuted event has three defined reason codes in section 3.17.4 of VAST 4.3: 1 for a rejected resource, 2 for unsupported verification and 3 for a resource-loading error. Those distinctions should survive your logging pipeline.
A single measurement-failed flag loses the information needed to decide whether to inspect a publisher policy, an execution environment or a resource fetch.
Keep the raw reason and the interpretation
Store the reported value before translating it into a label. If a producer sends an unfamiliar value, preserve it instead of silently assigning the closest familiar category.
This complete Node.js example formats a local diagnostic record. It does not execute verification code or fire any tracking URL:
import assert from 'node:assert/strict';
const reasons = new Map([
[1, 'resource rejected'],
[2, 'verification unsupported'],
[3, 'resource load error']
]);
function diagnostic(rawCode, context) {
const known = Number.isInteger(rawCode) && reasons.has(rawCode);
return {
raw_reason: rawCode ?? null,
reason_label: known ? reasons.get(rawCode) : 'unrecognized or absent',
reason_recognized: known,
vendor: context.vendor,
resource: context.resource,
source_document: context.sourceDocument,
player_build: context.playerBuild,
environment: context.environment,
spec: 'VAST 4.3 section 3.17.4'
};
}
const context = {
vendor: 'synthetic-vendor',
resource: 'https://vendor.example/verify.js',
sourceDocument: 'wrapper-hop-2.xml',
playerBuild: 'test-build-8',
environment: 'non-browser-js'
};
assert.equal(diagnostic(1, context).reason_label, 'resource rejected');
assert.equal(diagnostic(2, context).reason_label, 'verification unsupported');
assert.equal(diagnostic(3, context).reason_label, 'resource load error');
assert.equal(diagnostic(4, context).raw_reason, 4);
assert.equal(diagnostic(4, context).reason_recognized, false);
assert.equal(diagnostic('3', context).raw_reason, '3');
assert.equal(diagnostic('3', context).reason_recognized, false);
assert.equal(diagnostic(undefined, context).raw_reason, null);
The string case is deliberate. Parsing an incoming transport representation is a separate boundary. This formatter expects a validated numeric code and preserves evidence if its caller violates that contract.
Include a schema or event version for your diagnostic record as it evolves. Otherwise a downstream analytics job may interpret the same property differently after a logging refactor.
Investigate rejection before investigating connectivity
For reason 1, inspect the decision that prevented the resource from executing. Which vendor or domain policy applied? Which policy revision did the player use? Did the resource identity match what the integration expected?
A rejected resource may never have been fetched. Repeatedly checking whether its URL returns HTTP 200 from your laptop will not explain a policy decision made in the player.
Preserve the verification entry and the source document that introduced it. If several wrappers contribute vendor resources, a log containing only the final creative ID can identify the wrong party.
Keep any proposed policy change within the integration's actual approval process. The diagnostic goal is to explain why the resource was rejected, not to bypass a deliberate publisher control so a test appears green.
Investigate unsupported verification against the runtime
For reason 2, inspect the API framework, resource type and execution environment. Compare those declarations with the capabilities of the actual player and SDK build.
The browserOptional declaration is one relevant input when JavaScript resources are involved. A browser test harness can hide a dependency that a non-browser JavaScript runtime cannot satisfy.
Record the selected resource and the alternatives the player considered. A resource-selection error and a genuinely unsupported execution contract can produce similar missing-measurement symptoms but require different repairs.
Do not infer platform support from a marketing platform name or an operating-system label. Use the actual integration version and a controlled runtime test.
Investigate loading errors with request evidence
For reason 3, preserve the attempted URL, timing and observed fetch or loading failure. Check malformed URLs, unsuccessful responses and timeouts in the environment that made the request.
Separate what the player detected from your later diagnosis. A timeout is an observation. A vendor outage is a hypothesis until supporting evidence identifies it.
Keep the resource response revision when possible. A URL can remain unchanged while its script changes, so a later successful fetch may not reproduce the content used by the failing run.
Also retain local runtime diagnostics when available. The absence of a load-error report does not establish that every subsequent JavaScript operation succeeded. Some failures may not be observable through the same reporting mechanism.
Inspect the not-executed callback itself
The callback's [REASON] macro belongs to the verification-not-executed context. Preserve the template and expanded request in a controlled test so you can see whether the reported reason survived URI construction.
Distinguish three observations: the player decided verification was not executed, the player attempted its callback, and the receiver observed that callback. A missing callback at the receiver does not tell you which earlier step failed.
Likewise, silence is not a positive execution receipt. The URL might be absent, blocked or outside the instrumentation you captured. Keep an unknown outcome when the evidence cannot establish success or a specific failure.
Avoid manually replaying production tracking URLs during diagnosis. Local expansion tests and controlled test endpoints can inspect behavior without creating artificial measurement signals.
Turn the incident into a regression fixture
Keep separate fixtures for a rejected resource, an unsupported environment and a loading failure. For each, record the configured condition, selected verification entry, observed reason and callback behavior.
After the repair, repeat the affected fixture and a known working baseline. Preserve a positive measurement-session observation where your integration exposes one, alongside playback evidence.
I maintain VASTlint. A wrapper inspection can help locate the document that supplied a verification resource. Attach the player diagnostic record to that result so the incident points to a specific resource, environment and observed failure.
Sources
Technical sources checked October 25, 2026. Drafted with AI; examples are synthetic unless explicitly stated otherwise.
Top comments (0)