=== Foreversorted Capacity Projection ===
config_file=/home/shared/webprojects/operations/capacity/config/capacity.conf
foreversorted_metrics=/home/shared/webprojects/operations/capacity/output/foreversorted-20260506T090001Z/metrics.env
nextcloud_metrics=/home/shared/webprojects/operations/capacity/output/nextcloud-20260506T090008Z/metrics.env

Current observed (Foreversorted):
- peak_hour_rps=0.000556
- window_avg_rps=0.002222
- peak_minute_rps=0.033333
- peak_second_rps=1.000000
- total_requests_in_sample=2
- sample_lookback_minutes=15
- sample_start_ts=06/May/2026:18:54:35 +1000
- sample_end_ts=06/May/2026:18:54:37 +1000
- deploy_token_hits_in_sample=0
- avg_duration_us=0

Projection assumptions:
- current_user_count=9
- user_count_source=families_sessions_recent
- active_user_count=9
- total_user_count=19
- active_user_lookback_days=30
- observed_deploy_token_share=0.000000
- observed_app_reference_rps=0.002222
- per_user_app_reference_rps=0.000247
- deploy_token_poll_seconds=60
- deploy_token_active_ratio=1.0

Projected Foreversorted load:
- 100 users total_peak_rps=1.691367 (app_only_estimate=0.024700, token_estimate=1.666667)
- 1000 users total_peak_rps=16.913667 (app_only_estimate=0.247000, token_estimate=16.666667)

Nextcloud baseline:
- current_snapshot_peak_hour_rps=0.070278
- current_snapshot_window_avg_rps=0.282222
- current_snapshot_reference_rps=0.282222
- sample_lookback_minutes=15
- sample_start_ts=06/May/2026:18:45:12 +1000
- sample_end_ts=06/May/2026:19:00:00 +1000
- baseline_mode=rolling_median
- baseline_source=rolling_median_96_snapshots
- baseline_reference_rps=0.172222
- rolling_window_snapshots=96
- fixed_baseline_rps=not_set

Combined shared-host projection:
- 100 users combined_peak_rps=1.863589 (foreversorted_estimate=1.691367, nextcloud_baseline=0.172222)
- 1000 users combined_peak_rps=17.085889 (foreversorted_estimate=16.913667, nextcloud_baseline=0.172222)

Capacity planning gate (optional):
- configured_shared_host_rps_cap=25
- configured_foreversorted_rps_cap_legacy=25
- 100 users remaining_shared_host_headroom_rps=23.136411
- 1000 users remaining_shared_host_headroom_rps=7.914111
- 100 users approaches_80pct_shared_host_cap=no
- 1000 users approaches_80pct_shared_host_cap=no

Interpretation checklist:
- Use combined shared-host projection for server-upgrade timing while Nextcloud remains colocated.
- If Nextcloud moves to a separate server later, set its baseline to zero or disable the shared baseline.
- If token_estimate dominates, reduce token polling frequency/duplication first.
- Recalculate after separating /.deploy-token into its own access log.