Troubleshooting Center¶
Applies to: All subscriptions
Purpose¶
Route any reported problem to the right diagnostic path quickly, so first-line support resolves or correctly classifies most incidents without escalation — and every escalation that does happen arrives with the evidence already gathered.
Audience¶
Customer technical support, first-line and second-line. Administrators diagnosing their own environment.
Prerequisites¶
- Monitoring and Health Checks — the observability these paths depend on
- Access to the application, and to logs where the path calls for them
Steps¶
Start with triage¶
Problem reported
|
v
Is anyone able to use the system at all?
|
+-- No, nobody --------> Application, database or proxy down
| -> Performance and Availability
|
+-- No, one person ----> Sign-in, role, or registration access
| -> Login and Access Failures
|
+-- Yes, but ...
|
+-- documents are not arriving ------> Ingest Failures
|
+-- a document will not register ----> IRN Failures
|
+-- an e-Way Bill problem -----------> e-Way Bill Failures
|
+-- the ERP is missing outcomes -----> Write-back Failures
|
+-- everything is slow --------------> Performance and Availability
|
+-- a screen or module is missing ---> see the three-gates table below
The pages¶
| Page | Covers |
|---|---|
| Support Investigation | Start here when you have a document number, IRN, e-way bill number or request ID. One screen shows that document's full history, current status and who has to act |
| Decision Trees | The full trees, in one place, for printing or embedding in a runbook |
| IRN Failures | A document will not register |
| e-Way Bill Failures | Generation, Part-B, extension, cancellation |
| Ingest Failures | Documents not arriving, or arriving wrong |
| Write-back Failures | The ERP is out of step |
| Login and Access Failures | Cannot sign in, or cannot see what is expected |
| Performance and Availability | Slow, or down |
| Attention Center Issues | The Attention Center shows nothing, old figures, or an item that will not close |
The three gates — for "something is missing"¶
This resolves a large share of tickets on its own. The three access gates fail in visibly different ways:
| What the user sees | Gate | Fix |
|---|---|---|
| The whole screen or module is absent | Feature flag | Feature Flags |
| The screen is there, controls are greyed out | Role permission | Roles and Permissions |
| The screen loads and shows no data | Registration access | GSTIN Access Control |
Ask the reporter which of the three they are seeing. It is a single question and it usually ends the investigation.
Before you start any path¶
Gather these four. Every path assumes them, and gathering them first saves a second round of questions:
| Item | Why |
|---|---|
| What exactly is on screen | The literal message, not a paraphrase |
| Document number | Identifies the record |
| Request identifier | Turns log searching into a direct lookup |
| When, and whether it ever worked | Distinguishes a regression from a first attempt |
Classify before you diagnose¶
Most Complifly problems fall into four classes, and knowing the class points at the owner immediately:
| Class | Signature | Owner |
|---|---|---|
| Data | One document, a named field | The user |
| Configuration | Everything of one kind fails | An administrator |
| Statutory | A window, a cancelled registration, a consumed number | Nobody — it is the law |
| Platform | Errors unrelated to any document | Complifly support |
The statutory class is worth naming explicitly at triage. Roughly a third of what reaches support is a legal constraint, and no ticket can resolve one.
Validation¶
Support is working well when:
| Measure | Target |
|---|---|
| Triage time | Under a minute to the right path |
| First-line resolution | The majority, without escalation |
| Escalations carry evidence | Every one has the four items above |
| Statutory issues are not escalated | Routed to the tax lead instead |
| Repeat incidents fall | Systemic causes fixed, not instances handled |
| Trees are maintained | Every escalated incident is reviewed for whether a tree should have caught it |
Troubleshooting¶
| Symptom | Cause | Action |
|---|---|---|
| Every ticket is escalated | Trees not used, or not trusted | Review a sample; if a tree was wrong, fix the tree |
| Escalations lack evidence | The four items not gathered at intake | Make them mandatory fields on the ticket |
| The same problem recurs weekly | A systemic cause is being handled as instances | Find the cause. Instance handling will not keep up |
| A path does not fit the situation | The tree covers the common case | Escalate with what you have tried, and propose the tree change |
| Statutory issues escalated repeatedly | The classification is not being read | Route them to the tax lead at triage |
| A module is on or off for one registration and nobody knows why | A flag's value comes from the first of three places that has one: the GSTIN's own override, then the company default, then the platform default | Admin & Logs > Operational Controls > Effective View, for that GSTIN. It shows the value in force and which of the three set it, and links to where to change it. Audit History says who set it and why. See Feature Flags |
An organization administrator cannot change a setting marked Managed by Complifly (FEATURE_WRITE_NOT_PERMITTED) |
It is part of the subscription — only a Complifly Super Admin may change it | Expected. Raise it with Complifly; do not look for a role that grants it — none does |
| The company default was changed and one registration did not follow | That registration has its own override, which always wins | GSTIN Overrides — remove the override to make it follow the company |
| A feature change is refused naming a GSTIN that was not in the request | A company default reaches every GSTIN that inherits it; the refusal (APPROVALS_IN_FLIGHT or NON_ERP_POSTING_BLOCKED) is on that GSTIN's behalf |
Resolve it for the named GSTIN, or confirm the change with a written reason. See Platform Error Codes |
Related Articles¶
- Error Code Reference — what each code means
- Monitoring and Health Checks — catching problems before they are reported
- Operations Guide — the routine that prevents many of these
- FAQ — the questions behind the most common tickets