Platform Analytics

A real-time view of how your LoginRadius APIs are performing — traffic volume, latency percentiles, request breakdowns, and capacity peaks. Use it to investigate slow endpoints, locate the source of errors, and verify peak load stays within your subscribed RPS.

Access: Admin Console → Insights → Platform Analytics. If the feature is not enabled on your tenant, contact support@loginradius.com.

Looking for authentication health? See Security Analytics for login outcomes, failure reasons, password-reset, MFA, and account-protection signals.

---

Tabs

Pick a tab to slice traffic by that attribute. All charts total volume as a single line; every other tab fans out into one color-matched line per summary card so you can compare buckets side by side.

All

The default tab — total request volume with no breakdown. Use it as the big-picture starting point before slicing further.

19.68k

Total

19.64k(100%)

All requests

What to watch

• Sudden cliffs in traffic — usually a client disconnect or DNS issue.
• Sharp spikes — typically a marketing push, a client retry storm, or a runaway integration looping on errors.
• Once you spot something, switch to Status Code or API Path to investigate.

Status Code

Breaks request volume down by HTTP response code. The fastest way to spot if errors are climbing.

19.68k

Total

19.35k(98.35%)

200

279(1.42%)

403

36(0.18%)

302

8(0.04%)

400

What to watch

• 4xx climbing — client misuse or rate-limit hits. Cross-check IP Address & User Agent.
• 5xx appearing — server-side issue. Open a support ticket with the time window and affected paths.
• 302 increases — usually expected in redirect-heavy flows (OAuth).

HTTP Method

Breaks request volume down by HTTP verb. Useful for spotting unexpected method mixes.

19.68k

Total

18.06k(91.77%)

GET

1.60k(8.12%)

POST

15(0.08%)

PUT

6(0.03%)

DELETE

What to watch

• GET typically dominates for auth and profile reads. POST is smaller (sign-ups, logins). PUT/DELETE are usually under 1%.
• A sudden POST spike on a GET-heavy workload usually means a bulk import or misconfigured client — investigate via API Path.

API Path

Breaks request volume down by endpoint. The single highest-value tab when investigating performance.

19.68k

Total

13.30k(84.07%)

/identity/v2/auth/access_token/Validate

1.42k(8.96%)

/service/oauth/token

473(2.99%)

/v2/manage/2fa/config

328(2.07%)

/identity/v2/auth/access_token

What to watch

• Which endpoint is taking the brunt of a traffic surge?
• Which endpoint correlates with a 5xx spike on Status Code?
• To investigate one endpoint end-to-end, add filter API Path · Contains · /the-path.

User Agent

Breaks request volume down by the User-Agent header — which SDKs, browsers, and tools are calling your APIs.

19.68k

Total

11.31k(57.80%)

axios/1.16.1

7.50k(38.34%)

axios/1.15.2

507(2.59%)

Chrome

167(0.85%)

Mozilla/5.0

What to watch

• Outdated SDK versions — if a stale version still drives the bulk of traffic, schedule a client upgrade before deprecating server-side support.
• Unknown agents — non-LoginRadius, non-browser strings are worth investigating.
• The User Agent header is client-supplied and unverified — treat it as a strong hint about who is calling, never as proof of identity.

IP Address

Breaks request volume down by originating IP. First place to check for unusual traffic.

19.68k

Total

8.79k(46.27%)

3.214.251.65

6.44k(33.86%)

54.164.64.8

3.31k(17.41%)

14.192.1.101

260(1.37%)

49.37.157.237

What to watch

• A new IP in the top 5 that was not there last week — verify it is an expected partner or integration.
• A single IP responsible for > 50% of traffic — likely a server-side integration. Make sure it is authorised.
• If you spot a suspicious IP, add filter IP Address · Equals · <ip> to see what it is doing.

Host

Breaks request volume down by destination hostname. Useful in multi-environment or multi-region setups.

19.68k

Total

18.00k(91.50%)

api.example.com

1.42k(7.23%)

api-staging.example.com

250(1.27%)

accounts.example.com

What to watch

• Multi-environment — confirm dev/staging traffic is not hitting production hosts.
• Custom domains — verify a newly configured custom domain is receiving traffic.
• Migrations — watch traffic shift from an old hostname to a new one.

---

Latency Graph

Distribution of response latency over time, broken into four percentile bands — P75, P90, P95, P99. Each band is stacked so the gap between typical and worst-case performance is visible at a glance.

Unit

Milliseconds

Percentiles

P75 / P90 / P95 / P99

Scope

Respects filters

P75 LatencyP90 LatencyP95 LatencyP99 Latency

---

Request Volume by Source

A grid of cards, each showing the top five contributors for one attribute. Use it to find the loudest IPs, the busiest endpoints, the dominant SDK versions, and the response-code mix in one glance. Bar length is proportional to the largest row in the card.

Breakdown by

IP, Host, Method, Path, Status, UA

Visible rows

Top 5 · Load More extends in-page

CSV download

Top 100 per card

Each card lists the top 5 contributors. Bar length = relative volume; the largest is full width.

The cards at a glance

IP Address

Originating client IPs. A new IP in the top 5 is worth a quick look.

Host

Destination hostnames. Useful in multi-environment or multi-region setups.

Method

HTTP verb mix. Usually GET-dominant; a sudden POST surge is worth investigating.

Path

Endpoints by volume — the most common drill-down target.

Status

Response-code distribution. A growing 4xx/5xx share means trouble.

User Agent

SDKs and browsers hitting the API. Drift here can flag rogue integrations.

---

Requests Per Second (RPS)

Peak Requests-Per-Second across the selected range, charted day by day. Use it to understand the throughput your tenant is generating at peak.

Measures

Daily peak (not average)

Filters

Do not apply

Date picker

Date-only (no time)

Peak RPS for each day in the selected range.

Filters do not apply here. The page-level Add Filters control deliberately bypasses the RPS chart. RPS measures the load your tenant places on the platform, so it always reflects total traffic regardless of what you have filtered down to above.

---

Controls & Actions

Three controls scope the data and produce shareable output: Filters and Date Range at the top of the page, plus Export as PDF or per-card CSV.

Filters

Filters narrow every chart on the page to the slice of traffic you care about. They are global — apply once, every widget re-runs against them. The only exception is the RPS chart, which always reflects unfiltered traffic so capacity stays honest.

Scope

All charts except RPS

Rules

Combined with AND

Persistence

Cleared on reload

Anatomy of a filter

Each filter is one or more rules — attribute · operator · value. Click the + to add another rule; rules are combined with AND.

Clicking Select Rule opens the attribute list — pick one to start a filter.

Examples

Show only 5xx errors from auth endpoints

1. Open the Status Code tab.
2. Add filter: Status Code · Greater Than · 499.
3. Add filter: API Path · Contains · /auth.
4. Apply. The trend chart now shows only matching requests; Top 5 cards reveal the worst paths and IPs.

Hide health-check noise from a specific IP

1. Find the noisy IP in the Top 5 · IP Address card.
2. Add filter: IP Address · Not Equals · <that IP>.
3. Apply. Latency and trend charts now exclude that IP.

Investigate one client SDK across all endpoints

1. Add filter: User Agent · Contains · axios/1.15.
2. Open the API Path tab to see which endpoints that SDK version is hitting.

Date Range

The date range at the top right of the page bounds every trend chart and the latency graph. The RPS chart has its own independent picker so you can compare a long-term capacity view with a focused debugging window.

Minimum

15 minutes

Maximum

30 days

Default

Last 24 hours

Timezone

Account-configured

The date range field

The date range field at the top right of the dashboard. Click either input to open the picker.

Timezone

Every timestamp you see — picker inputs, chart axes, tooltip values, exported reports — uses your account's configured timezone (displayed as (UTC±HH:MM) next to the time inputs). To change it, update your account preferences in the Admin Console.

Export

Two ways to take data out of the dashboard: a page-wide PDF for what you are looking at right now, and per-card CSV for raw analysis.

Export as PDF

Click Export as PDF at the top right of the page. The export captures the entire view as currently shown — including the active filters, date range, every chart, and every Top 5 card. Active filters are stamped at the top of the PDF, which makes it suitable for audit trails and incident write-ups.

The Export as PDF button sits in the top right of the page header.

Download CSV (per Top 5 card)

Each Top 5 card has a download icon in its header. Clicking it exports the top 100 ranked rows for that attribute. Filters and the dashboard date range are applied before the file is generated.

Every card in the Top 5 grid has a download icon at the right of its header.

---

Best practices

Compare volume and latency side by side

Reading request volume next to latency is the quickest way to gauge how your tenant is performing under real load.

Watch Top IPs for anomalies

Glance at the Top IPs card regularly — a new or dominant IP is usually the first sign of a new integration, a scraper, or a misconfigured retry loop.

Use trend graphs to anticipate surges

Trend graphs surface the rhythm of your traffic — use them to anticipate surges and plan scaling before the queue builds up.

Review daily peak RPS values

Check the daily peak RPS to make sure your tenant is staying within the subscribed RPS capacity — and to spot the headroom you still have.

Leverage PDF exports

Snapshot the dashboard as a PDF for internal reviews, incident write-ups, or compliance reporting — filters and date range are stamped at the top.

Related Resources

Security AnalyticsUser AnalyticsAPI Rate LimitsError Response CodesAcceptable Use Policy