diff --git a/.tool-versions b/.tool-versions index c21f78518e..904bd54bc8 100644 --- a/.tool-versions +++ b/.tool-versions @@ -1,3 +1,3 @@ ruby 3.4.5 nodejs 22.19.0 -yarn 1.22.19 \ No newline at end of file +yarn 1.22.22 \ No newline at end of file diff --git a/src/data/nav/platform.ts b/src/data/nav/platform.ts index ce1a3b6682..5cbb4d0cc3 100644 --- a/src/data/nav/platform.ts +++ b/src/data/nav/platform.ts @@ -124,6 +124,27 @@ export default { link: '/docs/platform/pricing/limits', name: 'Limits', }, + { + name: 'Pricing examples', + pages: [ + { + link: '/docs/platform/pricing/examples/livestream', + name: 'Livestream chat', + }, + { + link: '/docs/platform/pricing/examples/support-chat', + name: 'Support chat', + }, + { + link: '/docs/platform/pricing/examples/data-broadcast', + name: 'Data broadcast', + }, + { + link: '/docs/platform/pricing/examples/realtime-dashboard', + name: 'Realtime dashboard', + }, + ], + }, { link: '/docs/platform/pricing/faqs', name: 'Pricing FAQs', diff --git a/src/pages/docs/platform/pricing/examples/data-broadcast.mdx b/src/pages/docs/platform/pricing/examples/data-broadcast.mdx new file mode 100644 index 0000000000..894c521e47 --- /dev/null +++ b/src/pages/docs/platform/pricing/examples/data-broadcast.mdx @@ -0,0 +1,89 @@ +--- +title: Data broadcast pricing example +meta_description: "A pricing example that uses Ably Pub/Sub for data broadcast. Example shows how message conflation reduces message costs from ~$1,800 to ~$370/month for 10K users across 50 matches." +meta_keywords: "sports betting, live odds, realtime odds updates, message conflation, Pub/Sub pricing, betting platform, live sports data, odds streaming, realtime data delivery, Ably Pub/Sub pricing" +intro: "This Pub/Sub example demonstrates consumption-based pricing for a realtime data broadcast – a single source publishing frequent updates to many subscribers." +--- +This example uses a sports betting scenario where odds are updated and fanned out to bettors. The same pattern applies to any broadcast where only the latest value matters, such as stock tickers, live scores, auction platforms and transport arrivals. + +## Assumptions + +- 10,000 monthly active users +- 50 live matches per month +- 200 concurrent viewers per match +- 2-hour match duration +- 60-minute average session duration +- 10 raw odds updates per second from data provider +- Message conflation enabled (500ms interval) + +These assumptions generate: +- [147,600,000 messages](#messages) +- [2,400,000 connection minutes](#minutes) +- [4,800,000 channel minutes](#minutes) + + + +## Cost summary + +The high-level cost breakdown for this scenario. Messages are billed for both inbound (published to Ably) and outbound (delivered to subscribers). + +| Item | Calculation | Cost | +|------|-------------|------| +| Messages (with conflation) | (3,600,000 inbound + 144,000,000 outbound = 147,600,000) × $2.50/M | $369 | +| Connection minutes | 2,400,000 × $1.00/M | $2.40 | +| Channel minutes | 4,800,000 × $1.00/M | $4.80 | +| Package fee | | [See plans](/docs/platform/pricing) | +| **Total** | | **~$376.20/month** | + +## Message breakdown + +How the message total is calculated. Conflation reduces outbound delivery from 10 updates/sec to 2 updates/sec (500ms interval). + +| Item | Calculation | Messages | +|------|-------------|------| +| Inbound (from data provider) | 10 updates/sec × 7,200 sec × 50 matches | 3,600,000 | +| Outbound (to viewers, with conflation) | 2 updates/sec × 7,200 sec × 200 viewers × 50 matches | 144,000,000 | +| **Total** | | **147,600,000** | + + +## Connection and channel minutes + +The following table shows how connection and channel minute costs are calculated in this example: + +| Metric | Calculation | Monthly | Cost | +|--------|-------------|---------|------| +| Connection minutes | 10,000 users × 4 sessions × 60 mins | 2,400,000 | $2.40 | +| Channel minutes | 40,000 sessions × 60 mins × 2 channels | 4,800,000 | $4.80 | + + +## Why conflation, not batching? + +For live betting, [message conflation](/docs/messages#conflation) is the right optimization because: + +- Old odds are semantically stale — batching would group outdated prices together +- Users need the latest price, not a history of price changes +- Conflation reduces both inbound and outbound message costs + +Use [server-side batching](/docs/messages/batch#server-side) instead when every message matters, for example in a chat use case or when you're sending push notifications. + + + +## Further optimization: Delta compression + +For richer payloads such as full market depth or live statistics, [delta compression](/docs/channels/options/deltas) can reduce bandwidth costs by sending only the difference between updates. + +| Payload type | Full size | With delta | Bandwidth reduction | +|--------------|-----------|------------|---------------------| +| Full market (10+ selections) | 8 KiB | ~2 KiB | 75% | +| Live stats + odds bundle | 15 KiB | ~4 KiB | 73% | + + \ No newline at end of file diff --git a/src/pages/docs/platform/pricing/examples/livestream.mdx b/src/pages/docs/platform/pricing/examples/livestream.mdx new file mode 100644 index 0000000000..b460128da5 --- /dev/null +++ b/src/pages/docs/platform/pricing/examples/livestream.mdx @@ -0,0 +1,77 @@ +--- +title: Major livestream event pricing example +meta_description: "Calculate Ably Chat pricing for livestream events with high-concurrency chat. Example shows 25K concurrent viewers, message batching reducing costs by 95%, and total cost of ~$2,430 for a 1-hour major event." +meta_keywords: "livestream chat, high concurrency chat, message batching, chat pricing, realtime messaging, room reactions, chat moderation, event chat, scalable chat, Ably Chat pricing" +intro: "This Ably Chat example demonstrates consumption-based pricing for a major live event – thousands of viewers chatting simultaneously during a broadcast. Livestream chat involves high message velocity over a short duration, where batching is essential to manage costs at scale." +--- + +## Assumptions + +The scale and features used in this calculation. + +| Scale | Features | +|-------|----------| +| 1-hour event duration | ✓ Message batching (100ms window) | +| 100 messages per second | ✓ Moderation (100% of messages) | +| 360,000 messages sent = 900,000,000 delivered (with batching) | ✓ Room reactions | +| 10,000 room reactions = 70,000,000 delivered (with batching) | | +| 25,000 viewers × 1 hour = 1,500,000 connection minutes | | + + + +## Cost summary + +The high-level cost breakdown for this scenario. Messages are billed for both inbound (published to Ably) and outbound (delivered to subscribers) – a single message to 100 subscribers generates 101 billable messages. + +| Item | Calculation | Cost | +|------|-------------|------| +| Messages (with batching) | 900,000,000 × $2.50/M | $2,250 | +| Room reactions (with batching) | 70,000,000 × $2.50/M | $175 | +| Connection minutes | 1,500,000 mins × $1.00/M mins | $1.50 | +| Moderation (Ably) | 360,000 msgs × 1 rule invocation | $1.00 | +| Package fee | | [See plans](/docs/platform/pricing) | +| **Total** | | **$2,427.50/event** | + +## Message breakdown + +How the message cost breaks down by type. + +| Type | Inbound | Outbound | Total | Cost | +|------|---------|----------|-------|------| +| Chat messages | 360,000 | 900,000,000 | 900,360,000 | $2,250 | +| Room reactions | 10,000 | 70,000,000 | 70.010,000 | $175 | +| Moderation* | 360,000 | — | 360,000 | $1.00 | +| **Message costs** | **730,000** | **970,000,000** | **970,730,000** | **$2,426** | + + + +## Batching impact + +How batching reduces costs at scale. Actual savings depend on your message patterns – bursty traffic batches more efficiently than steady streams. + +| Scenario | Cost | Messages | Savings | +|----------|------|----------|---------| +| Without batching | $22,500 | 9,000,000,000 messages | — | +| With batching | $2,250 | 900,000,000 messages | **$20,250** | + +## Room reactions calculation + +10,000 room reactions to 25,000 viewers would generate 250,000,000 outbound messages unbatched. + +With batching enabled and assuming 80% of reactions occur in bursts (during key moments), total outbound messages drop to 70,000,000: + +- **Burst reactions (80%):** 8,000 reactions batched into ~800 deliveries × 25,000 viewers = 20,000,000 messages +- **Individual reactions (20%):** 2,000 × 25,000 viewers = 50,000,000 messages + +**Total:** 70,000,000 messages × $2.50/M = **$175** + + \ No newline at end of file diff --git a/src/pages/docs/platform/pricing/examples/realtime-dashboard.mdx b/src/pages/docs/platform/pricing/examples/realtime-dashboard.mdx new file mode 100644 index 0000000000..44679181f7 --- /dev/null +++ b/src/pages/docs/platform/pricing/examples/realtime-dashboard.mdx @@ -0,0 +1,84 @@ +--- +title: Realtime dashboard pricing example +meta_description: "Calculate Pub/Sub pricing for healthcare patient monitoring dashboards. Example shows realtime vitals tracking for 100 patients monitored by 5 care coordinators, with total cost of ~$98/month including presence and history features." +meta_keywords: "healthcare dashboard, patient monitoring, realtime vitals, healthcare IoT, care coordination, patient monitoring devices, clinical dashboard, realtime healthcare data, Pub/Sub pricing, Ably healthcare" +intro: "This Pub/Sub example demonstrates consumption-based pricing for a realtime dashboard – many data sources publishing to a small number of viewers." +--- + +This example uses a healthcare scenario, sending patient vitals to care coordinators. The same pattern applies to any dashboard with high inbound volume and low fan-out, such as IoT sensor monitoring, logistics tracking, infrastructure observability, and more. + +## Assumptions + +The scale and features used in this calculation. + +| Scale | Features | +|-------|----------| +| 100 patients with home monitoring devices | ✓ Presence (shift handover visibility) | +| 5 care coordinators viewing dashboard | ✓ History retrieval (late joiners see recent readings) | +| 8-hour monitoring shifts, 22 days/month | | +| Vitals transmitted every 10 seconds (6 per minute) | | +| 15 clinical alerts per patient per day | | + +These assumptions generate: +- [38,210,000 messages](#messages) +- [1,110,000 connection minutes](#minutes) +- [1,110,000 channel minutes](#minutes) + +## Cost summary + +The high-level cost breakdown for this scenario. + +| Item | Calculation | Cost | +|------|-------------|------| +| Messages | 38,210,000 × $2.50/M | $95.53 | +| Connection minutes | 1,110,000 × $1.00/M | $1.11 | +| Channel minutes | 1,110,000 × $1.00/M | $1.11 | +| Presence & history | 6,600 messages | Less than $0.02 | +| Package fee | | [See plans](/docs/platform/pricing) | +| **Total** | | **$97.77/month** | + +## Message breakdown + +Patient devices transmit vitals (heart rate, SpO2, blood pressure) every 10 seconds, plus clinical alerts when readings breach thresholds. Each message is delivered to all 5 care coordinators subscribed to the monitoring channel. + +| Message type | Calculation | Monthly | +|--------------|-------------|---------| +| Vitals updates (inbound) | 100 patients × 6/min × 480 mins × 22 days | 6,340,000 | +| Clinical alerts (inbound) | 100 patients × 15/day × 22 days | 33,000 | +| **Total inbound** | | **6,370,000** | +| Outbound to care team | 6,370,000 × 5 coordinators | 31,850,000 | +| **Total messages** | | **38,210,000** | + +**Message cost:** 38,210,000 × $2.50/M = **$95.53** + +## Connection and channel minutes + +The following table shows how connection and channel minute costs are calculated in this example: + +| Metric | Calculation | Monthly | Cost | +|--------|-------------|---------|------| +| Patient device connections | 100 × 8 hrs × 22 days × 60 mins | 1,060,000 | $1.06 | +| Care coordinator connections | 5 × 8 hrs × 22 days × 60 mins | 52,800 | $0.05 | +| **Total connection minutes** | | **1,110,000** | **$1.11** | +| Channel minutes | 105 users × 8 hrs × 22 days × 60 mins | 1,110,000 | $1.11 | + +## Presence and history + +These features add negligible cost at this scale but provide important clinical functionality. Presence shows which coordinators are actively monitoring during shift changes, while history lets late joiners see recent readings. + +| Feature | Calculation | Monthly messages | +|---------|-------------|------------------| +| Presence events | 5 coordinators × 2 events × 22 days | 220 inbound | +| Presence fan-out | 220 × 4 other coordinators | 880 outbound | +| History retrieval | 5 coordinators × 1 request/shift × 22 days × 50 msgs | 5,500 | +| **Total** | | **6,600** | + +**Total cost:** 6,600 messages × $2.50/M = **$0.02** + + + diff --git a/src/pages/docs/platform/pricing/examples/support-chat.mdx b/src/pages/docs/platform/pricing/examples/support-chat.mdx new file mode 100644 index 0000000000..dc73d5c97b --- /dev/null +++ b/src/pages/docs/platform/pricing/examples/support-chat.mdx @@ -0,0 +1,79 @@ +--- +title: Enterprise support chat pricing example +meta_description: "Calculate Ably Chat pricing for enterprise customer support chat. Example shows 50K MAU, one-to-one messaging, and why consumption pricing at ~$78/month outperforms MAU pricing at $2,500/month for support use cases." +meta_keywords: "support chat, customer support, enterprise chat, one-to-one messaging, chat pricing, MAU pricing, consumption pricing, support agent chat, customer service chat, Ably Chat pricing" +intro: "This Ably Chat example demonstrates consumption-based pricing for enterprise customer support – one-to-one conversations between agents and customers. Support chat typically involves brief, infrequent sessions, making consumption-based pricing significantly more cost-effective than MAU pricing for this pattern." +--- + +## Assumptions + +The scale and features used in this calculation. + +| Scale | Features | +|-------|----------| +| 50,000 MAU (customers) | ✓ Typing indicators | +| 300 support agents | ✓ Presence | +| 5 conversations per customer/month (250,000 total) | ✓ History retrieval (20% of conversations, 50 messages each) | +| 20 messages per conversation | | +| 15 typing events per conversation | | +| 4 presence events per conversation | | +| 20-minute average conversation | | + +These assumptions generate: +- [27,000,000 messages](#messages) +- [5,530,000 connection minutes](#minutes) +- [5,000,000 channel minutes](#minutes) + +## Cost summary + +The high-level cost breakdown for this scenario. + +| Item | Calculation | Cost | +|------|-------------|------| +| Messages | 27,000,000 × $2.50/M | $67.50 | +| Connection minutes | 5,530,000 × $1.00/M | $5.53 | +| Channel minutes | 5,000,000 × $1.00/M | $5.00 | +| Package fee | | [See plans](/docs/platform/pricing) | +| **Total (Consumption)** | | **$78.03/month** | + + +## Message breakdown + +How the message cost breaks down by feature. Batching has minimal impact in 1:1 chat since there's only one recipient per message – batching benefits scale with room size. + +| Feature | Events | Inbound | Outbound | Total messages | +|---------|--------|---------|----------|----------------| +| Chat messages | 5,000,000 | 5,000,000 | 10,000,000 | 15,000,000 | +| Typing indicators | 3,750,000 | 3,750,000 | 3,750,000 | 7,500,000 | +| Presence | 1,000,000 | 1,000,000 | 1,000,000 | 2,000,000 | +| History retrieval | 50,000 requests | — | 2,500,000 | 2,500,000 | +| **Total** | | | | **27,000,000** | + +**Message cost:** 27,000,000 × $2.50/M = **$67.50** + +## Connection and channel minutes + +The following table shows how connection and channel minute costs are calculated in this example: + +| Type | Calculation | Minutes | Cost | +|------|-------------|---------|------| +| Customer connections | 250,000 conversations × 20 mins | 5,000,000 | $5.00 | +| Agent connections | 300 agents × avg 80 mins/day × 22 days | 528,000 | $0.53 | +| Channel minutes | 250,000 conversations × 20 mins | 5,000,000 | $5.00 | + +### Consumption vs MAU comparison + +Consumption-based pricing works better for support chat because customers connect briefly and infrequently, using far less than the Monthly Active User (MAU) allowances which are 20,000 messages and 2,000 connection minutes. + +| Model | Calculation | Monthly cost | +|-------|-------------|--------------| +| Consumption-based | As above | ~$78 | +| MAU pricing | 50,000 MAU × $0.05 | $2,500 | + + +