Tags, entry points and collections

Structure is what turns a long list of scenarios into answers. With a little consistent organisation you can say in seconds what to run before a release, where the risk sits, and which promises are non-negotiable. Without it, the list is a pile.

Four things do the work: entry points, tags, collections and red lines. Decide each once, write it down, and keep it consistent.

What to decide, and who decides

Decision	Recommended owner	Why it matters
The tag vocabulary	QA or engineering lead	Inconsistent tags make filtering unreliable.
Which collections exist and what each is for	Engineering lead with product	These are your release gates.
Which scenarios are red lines	Leadership	This is your risk appetite, written down.
Entry point naming	QA lead	It drives coverage reporting.

Entry points

The Entry point says where a scenario applies. It groups the list, feeds the area charts and drives Route coverage on Insights.

• Use your app’s real paths: /checkout, /login, /account/orders. Use / for something site-wide.
• Be consistent. /checkout and /Checkout page count as two different areas.
• One entry point per scenario. If it spans pages, choose where it matters most.

Qarmiq matches entry points to the pages it finds when it crawls your environment. Consistent paths keep the coverage number honest. See Environments.

Scenario tags

Tags are free labels for slicing the list. Keep a small, agreed set.

Kind	Examples	Answers
Importance	critical	What matters most?
Purpose	regression, smoke, security	What do we run for this reason?
Business area	payments, search, onboarding	Which team or domain owns it?

Good habits:

• Lowercase and one word where you can, such as payments.
• Five tags used consistently beat fifty used once.
• Do not repeat what Qarmiq already shows, such as the entry point or the health state.
• A scenario’s tags are separate from a check’s type. See Write good checks.

Collections: your release gates

A collection is a named set of scenarios for one purpose. Name it for the purpose, not the contents.

Collection	What goes in it	Run it
Smoke	The few scenarios that tell you in minutes whether a build is alive.	On every build.
Regression	Everything that must still hold after a change.	Before releases.
Release candidate	What you require before shipping.	Before each release.
Security	The scenarios about access and privacy.	On a schedule you set.

Why this is worth doing:

• A run can be created from a collection, so the right set is one action away. See Create and find runs.
• Each collection shows on Insights with clean runs in a row. A release gate that has gone quiet, or stopped being clean, is visible without anyone reporting it.
• A scenario can belong to several collections.

Red lines: what you will not accept breaking

A red line is a scenario that must never fail. If one fails, you are told at once, and again when it holds. See History and red lines.

• Keep the list short. If everything is a red line, none is.
• Protect money, data and trust: payment works, one customer never sees another’s data, sign-in refuses the wrong person.
• Name who holds each one, and add the teammates who should hear about a breach.
• Review the list whenever the product or the business changes.

A worked example

Scenario	Entry point	Tags	Collections	Red line
Shoppers should always be able to complete checkout with a valid card.	/checkout	critical, payments	Smoke, Regression, Release candidate	Yes
The cart should always show the correct total.	/cart	regression, payments	Regression	No
Product search should always return relevant results.	/search	search	Regression	No
A shopper should never see another shopper’s order.	/account/orders	critical, security	Regression, Security	Yes

Review the structure

Use the list filters to inspect it: Health for what is failing or stale, Red lines for what must never fail, Collections for what a release needs. Merge duplicates as you find them. See Merge and collections. When you rename a tag or an entry point, change it everywhere in one pass.

Questions to ask your team

• What exactly must be green before we release, and is it a collection?
• Which scenarios are red lines, and who is told when one breaks?
• Does our coverage number reflect reality, or do our entry points not match our routes?
• Do two teams use different tags for the same thing?