Slice Time‑Out Parameters and Impact (Arc vs TM1 Configuration)

Slice Time‑Out Parameters and Impact

This page explains how time‑outs affect Slice users, what parameters control them across Arc and TM1, and how to tune settings to reduce unexpected disconnects without compromising server stability.

Client/Arc Layer

Slice “Arc Session Time Out” must be ≤ or matched to Arc’s sessiontimeout; otherwise UI may display a longer period than Arc permits.
Arc timeoutseconds must be long enough for the largest expected REST call (e.g., large view export).
Using Batch Requests reduces number of calls, lowering compounded timeout risk.

Server/TM1 Layer

ClientSessionTimeout should exceed typical user idle windows to avoid drops during analysis.
Idle HTTP connection timeouts set too aggressively will terminate long‑lived queries or queued requests.
Faster cubes (threading, feeders, rule optimization) mean fewer long‑running calls → fewer timeouts.

4) Recommended Baselines

Arc settings.yml
- sessiontimeout: 30–60 minutes for general analyst use.
- timeoutseconds: 120–300 seconds for on‑prem; 300–600 seconds for cloud/remote networks.
Slice → Settings → Connection
- Set “Arc Session Time Out” to match Arc sessiontimeout.
- Enable Batch Requests for workbooks with many DBRW/DBR formulas.
- Enable Use Async Requests on IBM Cloud; consider enabling on-prem only if frequent timeouts occur.
tm1s.cfg (TM1)
- ClientSessionTimeout: 60–120 minutes where users take breaks between analysis cycles.
- Idle HTTP/REST timeouts: Keep comfortably above Arc timeoutseconds to avoid mid‑request termination.

5) Can We Increase Time‑Out Automatically?

There are two complementary approaches:

Configuration‑based (preferred): Set higher, sensible defaults centrally in Arc settings.yml and in tm1s.cfg, then distribute Slice client settings. This is predictable and auditable.
Behavioral: Use Slice’s Async Requests and Batch Requests to reduce the chance of hitting request timeouts without changing server limits.

What you can “automate”: centrally manage Arc configuration and distribute standardized Slice settings. For variable needs (e.g., power users), create separate Arc connection entries with higher timeoutseconds and label them clearly.

6) Impact Analysis

Raising Arc sessiontimeout: Fewer browser session drops; slightly longer retention of idle sessions in Arc memory.
Raising Arc timeoutseconds: More long‑running queries can finish; hung calls also take longer to fail (monitor logs).
Raising TM1 ClientSessionTimeout: Fewer forced logouts; potential increase in concurrent sessions during peak periods.
Relaxing idle REST timeouts: Improves reliability for slow networks; keep safeguards via reverse proxies/WAF where applicable.
Using Batch/Async in Slice: Reduces timeout incidence without changing server config; may alter user feedback timing (requests complete in the background).

7) Step‑by‑Step Tuning Playbook

Baseline today’s behavior: collect examples of time‑out errors, affected actions, and durations.
Set Arc sessiontimeout=45–60 mins and align Slice “Arc Session Time Out”.
Increase Arc timeoutseconds to 300–600s for heavy reports or remote users.
Enable Slice Batch Requests; enable Async Requests for IBM Cloud.
On TM1, review ClientSessionTimeout and idle REST timeouts; ensure they exceed Arc’s values.
Test representative workbooks; observe improvements and any side effects (resource usage, queue times).
Monitor with Pulse and Arc logs; adjust incrementally.

8) Examples and Templates

8.1 Arc settings.yml snippets

8.2 Slice client guidance

- Arc Session Time Out: 60
- Batch Requests: Enabled
- Use Async Requests: Enabled for IBM Cloud; conditional for on‑prem

8.3 tm1s.cfg checklist

Confirm HTTPPortNumber is defined and open.
Set ClientSessionTimeout ≥ 60.
Review idle REST/HTTP timeout parameters; keep above Arc timeoutseconds.
Validate after restart using Arc and representative Slice workbooks.