Grafana in space: Monitoring Japan's SLIM moon lander in real time

Presented at GrafanaCON2024.
https://grafana.com/about/events/grafanacon/2024/

Find original movies through QR codes or links below,
https://www.isas.jaxa.jp/home/slim/SLIM/
https://darts.isas.jaxa.jp/

1. NAKAHIRA, Satoshi (ISAS/JAXA) on behalf of YOKOTA, Kentaro ; ITO, Takahiro; GOTO, Kenta; AKIZUKI, Yuki; KANAYA, Shuusaku ; and SAWAI, Shujiro Grafana in space: Monitoring Japan's SLIM moon lander in real time

2. Agenda for today’s talk About me About ISAS/JAXA Overview and Results of SLIM Project How Grafana is used in SLIM operations Our requirements for new “Quick-Look (QL) System” System Overview Examples of Grafana dashboards we Created

3. About me Majoring in Astrophysics (Ph.D.) Highlight of my work as an astronomer: Soft X-ray all-sky map with MAXI/SSC. doi : 10.1093/pasj/psz139 My focus is shifting to engineering, Open data archiving of space science observational data. (DARTS; https://darts.isas.jaxa.jp/ ) Operational systems of space science satellites, probes. Working on “ Science Satellite Operation and Data Archive Unit” at ISAS/JAXA

4. About ISAS/JAXA Akatsuki: Venus orbiter HINODE: Solar Telescope KAGUYA : Moon orbiter  Hayabusa2:  Sample return from Asteroids  ISAS Mission List Visit DARTS to get open science data https://darts.isas.jaxa.jp/

5. Demonstration of high-precision lunar landing technology The goal is a pin-point landing with 100 m accuracy. A chieved through "Image matching navigation" and "autonomous navigation guidance control" Realization of a lightweight explorer system 210 kg (body) + 500 kg (propellant) Missio n of SLIM SLIM landing site   © NASA/LRO © 藤井大地 @dfuji1, LRO https://twitter.com/i/status/1748103951336227113 “ Shioli ” crater SLIM Twitter Project Site

6. EARTH 2023.09.07 Launch 2023.10.01 Trans-Lunar Moon Orbit 2023.09.07 Launch (H-IIA 47; SLIM and XRISM) 2023.09.22 Navigation Camera Checkout Journey of SLIM

7. EARTH 2023.10.04 Swing-by Moon Orbit 2024.01.19-20 Descending & Landing 2023.12.25 Lunar orbit Intersection (LOI) 2023.12.25 LOI to 2024.01.19 Descend Journey of SLIM Farthest point >3x as far as Moon

8. LEV-1 LEV-2 (SORA-Q) Released two robots just before landing. Soft landing (the first in Japan). Pinpoint landing within 100 meters accuracy (the first in the world). Selfie taken by using LEV-2. Survive two extreme temperature shifts from night (-170 ℃ ) to day (110 ℃ ) it woke up at 2/ 25 and 3/ 27 . Highlights of SLIM

9. How Grafana is used in SLIM operations

10. Classic Quick-Look System and problems Classic Quick-Look System Built as monolithic software. Not easy to maintaining or adding features. Too many characters! Not easy to find a certain parameter Depends on human familiarity

11. How is telemetry on a scientific probe? Voltage Current Temperature Attitude Position Velocity A cceleration Status Flags Counts HK (House Keeping) data Science data: Image Event by event data of photon or particle Spectra Spectrogram Bursted time series data etc .. Monitor the health and functionality Usually analyzed on a slower time scale. Basic metrics, similar to what you find in IoT devices. (Not focused in the current system.)

12. IoTs: A small amount of telemetry from many of devices Probes: Large number of telemetries from one device.   How is the science probe’s telemet ry like Inspired by IoTs , we integrate Grafana to space p robe’s operations for enhanced Observability .

13. communication protocol handling Binary to values conversion C libs for both are InfluxDB Client Data receive/resister module Grafana InfluxDB (Python script) API (python) REST API REST API REST API REST API red text = newly written, other = OSS or reused software Wrapped with Cython configured in a loosely coupled manner based on REST API System Configuration Data calculation Image generation SLIM control Center import Python library for satellite specific process,

14. Grafana in the control room Inside the control room during the landing on the moon. Control room and YouTube audience, were all looking at the same dashboard at the same time.

15. Over 20 dashboards were constructed for each objective and operational phase. Classic QLs still dominant, but Grafana-based system played important role Examples of SLIM dashboards

16. stats timeseries logs XY plot Dashboards: Navigation system during the Cruise Phase Error/ Warning mode Messages

17. timeseries state timeline Gauge Dashboards: Heater Status Power-Related Metrics Heater Power temperature Heater ON/OFF Scrollable

18. (panel) Backend: Python Fast API endpoint Offline rendering with trimesh+pyrender Dashboards: thrusters and maneuvers (data source) Infinity REST API V olkovlabs Base64 Bar chart Gauge Sun - view Earth- view This video is available on:

19. import " influxdata / influxdb /schema" schema.fieldKeys (bucket: v.bucket , start: v.timeRangeStart , stop: v.timeRangeStop ) |> filter( fn : (r) => r._value =~ /${ filter_string }/ ) filter_string = variable1-textbox variable2-query The "variable" has worked very efficiently for us. Select the telemetry to be displayed. Easily locate and display a specific telemetry among thousands. Dashboards: querying with telemetry name How it's set up:

20. This video is available on: Actual data during the Moon landing Dashboards: Landing 32x

21. 8x Actual data during the Moon landing Dashboards: Landing This video is available on:

22. We were able to build the entire system quite easily. The visualization capabilities were broad, ranging from simple and clear to detailed graphs. We were also able to create dashboards that were well-received and visually impressive. Editing was straightforward and could be done even during operations. The "variable" feature was a game-changer and extremely convenient. Being able to call Python from the API meant that we could do anything. We’re glad to use Grafana . Conclusion / our impression to Grafana

23. Thank you very much, Grafana community! Let's go to the next planet together with JAXA!