Skip to content

Commercial

What we commit to, and what we will not put a number on yet

Support hours, a severity ladder, response and resolution targets, who you escalate to and what is excluded. Every target below is labelled, and one number is deliberately absent — the page says which, and why, rather than leaving you to notice.

The short version, if you are repeating this to somebody

Something breaks, you reach an on-call engineer directly, 24/7 — not a queue that opens in the morning. A critical incident gets a first response inside 15 minutes and an update every 30 until there is a workaround. If it passes that without one, the engagement lead is in it; the step after that is the owner of the business, by name, in your agreement. Every target on this page is indicative until it is written into that agreement, and the one number we publish nowhere is an availability percentage — the reason is below and it is not a hedge.

The honest shape of a service level

A service level is three separate promises that vendors habitually blur into one: when somebody answers, how fast a problem is worked, and what happens commercially if neither holds. This page separates them, because the first two are operational and the third is contractual, and only the third belongs in an agreement you sign.

Support is 24/7. That is a commitment rather than an aspiration, it is given in Asia/Dhaka time as the working hub, and it is the one figure on this page that is not qualified.

The response and resolution targets below are labelled indicative. They are the starting position for an agreement rather than a published guarantee, and they are written in full instead of withheld because a buyer cannot evaluate a vendor against a blank. What makes them honest is the label, and the label is a column in the table rather than a footnote under it.

The uptime percentage is not on this page, on purpose

We publish no availability figure at all — not the two or three decimal places every vendor page carries, not any number. Two reasons, and both are disqualifying on their own. Nothing in the platform measures availability continuously today, so any figure would be an assertion rather than a measurement. And an uptime percentage is a contractual liability with a credit schedule attached, which is a commitment the owner of this business signs rather than something a website decides. Availability is agreed per contract, against your own architecture, and a vendor quoting you a percentage before seeing your deployment is quoting you somebody else's. A number kept off this page cannot be screenshotted, quoted back, or picked up by a crawler as a commitment we never made.

Specification

Four severities, defined by impact rather than by feeling

Severity is assigned from what the viewer experiences, not from how the reporter feels about it. That is the only definition that survives a bad night.

LevelWhat it meansExample
S1 — CriticalService down, or playback failing for most viewersThe edge is not serving segments; sign-in is refusing everyone
S2 — MajorA core function is broken for a significant groupPayments failing at checkout; one live channel dark
S3 — MinorSomething is wrong and the service is usableA category page mis-sorting; a metadata field not saving
S4 — RequestA question, a change, or a scheduled piece of workA new plan configured; a ladder re-tuned; an ad tag swapped

Specification

Response and resolution

Every row carries its own label. A target without one is a commitment, and none of these has been signed yet.

LevelFirst responseOngoing updatesResolution aimStatus
S1 — Critical15 minutesEvery 30 minutes4 hours, or a workaround in placeIndicative — confirmed in the agreement
S2 — Major1 hourEvery 4 hours1 business dayIndicative — confirmed in the agreement
S3 — Minor1 business dayOn changeNext scheduled releaseIndicative — confirmed in the agreement
S4 — Request2 business daysOn changeScheduled with youIndicative — confirmed in the agreement
Support availability24/7Committed — 24/7, Asia/Dhaka hub
Platform availabilityNot publishedAgreed per contract — no figure is published
Resolution is an aim rather than a promise for a reason worth stating: a fault whose cause is in a third party — a gateway, a transit provider, a store review queue — is worked continuously and cannot be resolved to a clock we do not control.

Process

Who you reach, in order

An escalation path is only real if the last step is a person with a name, and it is.

01

The on-call engineer

Reached directly, 24/7, through the channel agreed at onboarding — not through a ticket queue that opens in the morning. This is where an S1 starts.

02

The engagement lead

Brought in when an S1 passes its update interval without a workaround, or when an S2 is going to miss its day. They can commit people rather than only report.

03

The owner

Named in the agreement. A small company's real escalation advantage is that the last step is one step away, and it is dishonest to draw four layers that do not exist.

What is outside it, said before you need it

Anything upstream of us. A payment gateway that is down, a store review that is slow, an ISP routing problem between a viewer and the edge, a content licence dispute — these are worked and reported and they cannot be resolved to our clock, because the fix is somebody else's.

Changes you make. On an owned platform the estate is yours, and if your team changes infrastructure outside the agreed process, the incident that follows is a support case rather than an SLA breach. That is the trade that comes with holding the keys, and it is a fair one.

Capacity you have not bought. If the deployment is sized for the peak you specified and the peak triples on the night of a match, the platform behaving like a platform at its limit is not a failure of the agreement. Sizing for events is a discovery conversation and it is cheaper than the incident.

Scheduled maintenance, agreed in advance with a window you approve. It is excluded because it is planned with you, not because it is convenient to exclude.

Questions

Questions

Why will you not publish an uptime percentage?

Because nothing measures availability continuously today, so the figure would be an assertion rather than a measurement — and because it is a contractual liability with a credit schedule behind it, which is a decision the owner signs rather than a page states. Availability is agreed per contract. A vendor who quotes you a percentage before seeing your deployment is quoting you somebody else's number.

What does 'indicative' actually mean here?

That the target is our starting position and becomes binding when it is written into your agreement. It is not a hedge that survives signature: whatever you and we agree is what we are held to. Publishing them unsigned is more useful than publishing nothing, provided the label never comes off.

Is 24/7 support real at your size?

It is a commitment given by the owner of the business and it is the one figure on this page carrying no qualifier. The honest detail behind it is that the escalation path is three steps rather than a tiered call centre, which is a disadvantage on volume and an advantage on the night something is actually wrong.

What if we operate the platform ourselves?

Then the service level covers the platform and the engineering support around it, not the operation of your estate. That split is written explicitly into the agreement, because an SLA over infrastructure somebody else administers is the kind of clause that is only discovered during an incident.

Bring your own SLA and we will mark it up

Most operators arrive with terms from a previous vendor or from a regulator. Send them and you will get back which clauses we can sign as they stand and which need a conversation.